Integrated Science Assessment
for Particulate Matter
First External Review Draft
ISA: EPA/600/R-08/139
Annexes: EPA/600/R-08/139A
National Center for Environmental Assessment-RTP Division
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC
-------
Disclaimer
This document is the first external review draft for review purposes only and does not constitute U.S.
Environmental Protection Agency policy. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
-------
Table of Contents
List of Tables xii
List of Figures xv
PM ISA Project Team xxm
Authors, Contributors, Reviewers xxvi
Clean Air Scientific Advisory Committee for Particulate Matter NAAQS xxxi
Chapter 1. Introduction 1 -1
1.1. Legislative Requirements 1-2
1.2. History of Reviews of the NAAQS for PM 1 -4
1.3. Document Development 1-9
1.4. Document Organization 1-10
1.5. EPA Framework for Causal Determination 1-11
1.5.1. Scientific Evidence Used in Establishing Causality 1-12
1.5.2. Association and Causation 1-13
1.5.3. Evaluation of Evidence for Going beyond Association to Causation 1-13
1.5.4. Application of Framework for Causal Determination 1-18
1.5.5. First Step—Determination of Causality 1-20
1.5.6. Second Step—Evaluation of Response 1-22
1.5.7. Concepts in Evaluating Adversity of Health Effects 1-24
1.6. Summary 1-25
Chapter 2. Integrative Health Effects Overview 2-1
2.1. Concentrations and Sources of Atmospheric PM 2-2
2.1.1. Ambient PM Variability and Correlations 2-2
2.1.1.1. Spatial Variability across the U.S. 2-2
2.1.1.2. Spatial Variability on the Urban and Neighborhood Scales 2-4
2.1.2. Temporal Variability 2-5
2.1.3. Correlations between Copollutants 2-5
2.1.4. Measurement Techniques 2-6
2.1.5. PM Source Characteristics 2-6
2.1.6. Source Contributions to PM 2-7
2.1.7. Policy-Relevant Background 2-8
2.2. Human Exposure 2-8
2.2.1. Outdoor Exposure to Ambient PM 2-8
2.2.2. Indoor and Personal Exposure to Ambient PM 2-9
2.2.3. Implications for Epidemiologic Studies 2-10
2.3. Health Effects 2-11
2.3.1. Exposure to PM10 2-12
2.3.1.1. Effects of Short-Term Exposure to PM10 2-12
2.3.1.2. Effects of Long-Term Exposure to PM10 2-14
2.3.2. Exposure to PM2.5 2-15
2.3.2.1. Effects of Short-Term Exposure to PM2.5 2-15
2.3.2.2. Effects of Long-Term Exposure to PM2.5 2-18
2.3.3. PM2.5 Constituents or Sources Linked to Health Outcomes 2-20
December 2008
DRAFT-DO NOT QUOTE OR CITE
-------
2.3.4. Public Health Impacts 2-23
2.3.4.1. PM Concentration-Response Relationship 2-23
2.3.4.2. Potentially Susceptible and Vulnerable Subpopulations 2-24
Chapter 3. Source to Human Exposure 3-1
3.1. Introduction 3-1
3.2. Overview of Basic Aerosol Properties 3-3
3.3. Sources of Primary and Secondary PM 3-6
3.3.1. Emissions of Primary PM and Precursors to Secondary PM 3-9
3.3.2. Formation of Secondary PM 3-12
3.3.2.1. Formation of Nitrate and Sulfate 3-12
3.3.2.2. Formation of Secondary Organic Aerosol 3-12
3.4. Monitoring Issues 3-14
3.4.1. Ambient Measurement Techniques 3-14
3.4.1.1. Federal Reference Method and Federal Equivalent Method Evaluation 3-14
3.4.1.2. PM Speciation 3-17
3.4.1.3. Ultrafine PM and PM Size Distribution 3-24
3.4.1.4. Multiple-Component Measurements on Individual Particles 3-25
3.4.1.5. Emerging Methods 3-26
3.4.2. Ambient Network Design 3-26
3.4.2.1. Monitor Siting Requirements 3-26
3.4.2.2. Spatial and Temporal Coverage 3-28
3.5. Ambient PM Concentrations 3-38
3.5.1. Spatial Distribution 3-38
3.5.1.1. Variability across the U.S. 3-39
3.5.1.2. Urban-Scale Variability 3-58
3.5.1.3. Neighborhood-Scale Variability 3-82
3.5.2. Temporal Variability 3-88
3.5.2.1. Trends 3-88
3.5.2.2. Seasonal Variations 3-94
3.5.2.3. Hourly Variability 3-96
3.5.3. Statistical Associations with Copollutants 3-100
3.5.4. Estimating Source Contributions to PM 3-103
3.5.4.1. Receptor Models 3-103
3.6. Background PM 3-115
3.6.1. Contributors to PRB levels of PM 3-115
3.6.1.1. Estimating PRB Concentrations 3-116
3.6.1.2. CTM for Predicting PRB Concentrations 3-118
3.7. Issues in Exposure Assessment for PM and its Components 3-129
3.7.1. Introduction and Key Concepts 3-129
3.7.2. Methods for Estimating PM Exposures 3-133
3.7.2.1. Exposure Monitoring and Associated Instrumental Measurement Errors 3-133
3.7.2.2. Uncertainties in PM Exposure Assessment 3-135
3.7.2.3. PM Exposure Modeling 3-139
3.7.3. Findings from PM Exposure Studies 3-143
3.7.3.1. Outdoor Exposure to Ambient PM 3-143
3.7.3.2. Indoor and Average Personal Exposure to Ambient and Non-Ambient PM 3-147
3.7.4. Exposure Assessment and Socioeconomic Status 3-161
3.8. Summary and Conclusions 3-167
3.8.1. Concentrations and Sources of Atmospheric PM 3-167
3.8.1.1. Ambient PM Variability and Correlations 3-167
3.8.1.2. Temporal Variability 3-170
3.8.1.3. Correlations between Copollutants 3-170
3.8.1.4. Measurement Techniques 3-171
3.8.1.5. PM Source Characteristics 3-171
3.8.1.6. Source Contributions to PM 3-171
3.8.1.7. Policy-Relevant Background 3-172
December 2008
iv
DRAFT-DO NOT QUOTE OR CITE
-------
3.8.2. Human Exposure 3-173
3.8.2.1. Outdoor Exposure to Ambient PM 3-173
3.8.2.2. Indoor and Personal Exposure to Ambient PM 3-173
3.8.2.3. Implications for Epidemiologic Studies 3-174
Chapter 4. Dosimetry 4-1
4.1. Introduction 4-1
4.1.1. Size Characterization of Inhaled Particles 4-1
4.1.2. Structure of the Respiratory Tract 4-2
4.2. Particle Deposition 4-5
4.2.1. Mechanisms of Deposition 4-6
4.2.2. Deposition Patterns 4-8
4.2.2.1. Total Respiratory Tract Deposition 4-9
4.2.2.2. Extrathoracic Region 4-10
4.2.2.3. Tracheobronchial and Alveolar Region 4-11
4.2.2.4. Localized Deposition Sites 4-12
4.2.3. Interspecies Patterns of Deposition 4-13
4.2.4. Biological Factors Modulating Deposition 4-14
4.2.4.1. Age 4-14
4.2.4.2. Gender 4-15
4.2.4.3. Anatomical Variability 4-16
4.2.4.4. Respiratory Tract Disease 4-17
4.2.4.5. Hygroscopicity of Aerosols 4-18
4.2.5. Summary 4-19
4.3. Clearance of Poorly Soluble Particles 4-20
4.3.1. Clearance Mechanisms and Kinetics 4-20
4.3.1.1. Extrathoracic Region 4-21
4.3.1.2. Tracheobronchial Region 4-21
4.3.1.3. Alveolar Region 4-22
4.3.2. Interspecies Patterns of Clearance and Retention 4-22
4.3.3. Particle Translocation 4-24
4.3.3.1. Alveolar Region 4-24
4.3.3.2. Olfactory Region 4-26
4.3.4. Factors Modulating Clearance 4-28
4.3.4.1. Age 4-28
4.3.4.2. Gender 4-29
4.3.4.3. Respiratory Tract Disease 4-29
4.3.4.4. Particle Overload 4-30
4.3.5. Summary 4-31
4.4. Clearance of Soluble Materials 4-32
4.4.1. Clearance Mechanisms and Kinetics 4-32
4.4.2. Factors Modulating Clearance 4-33
4.4.2.1. Age 4-33
4.4.2.2. Exercise 4-34
4.4.2.3. Disease 4-34
4.4.2.4. Concurrent Exposures 4-35
4.4.3. Summary 4-36
Chapter 5. Possible Pathways/ Modes of Action 5-1
5.1. Pulmonary Effects 5-2
5.1.1. Reactive Oxygen Species 5-2
5.1.2. Activation of Cell Signaling Pathways 5-4
5.1.3. Inflammation 5-5
5.1.4. Epithelial Barrier Function 5-5
5.1.5. Antioxidant Defenses and Adaptive Responses 5-6
5.1.6. Pulmonary Function 5-7
5.1.7. Allergic Disorders 5-8
5.1.8. Impaired Lung Defense Mechanisms 5-8
December 2008
v
DRAFT-DO NOT QUOTE OR CITE
-------
5.1.9. Resolution of Inflammation/Progression of Disease 5-9
5.1.10. Pulmonary DNA Damage 5-9
5.2. Systemic Inflammation 5-9
5.2.1. Endothelial Dysfunction and Altered Vasoreactivity 5-11
5.2.2. Activation of Coagulation and Acute Phase Response 5-12
5.3. Activation of the Autonomic Nervous System by Pulmonary Reflexes 5-13
5.4. Translocation of Ultrafine PM or Soluble PM Components 5-14
5.5. Disease of the Cardiovascular and Other Organ Systems 5-15
5.6. Results of New Inhalation Studies which Contribute to Modes of Action 5-15
Chapter 6. Integrated Health Effects of Short-Term PM Exposure 6-1
6.1. Introduction 6-1
6.1.1. Methodological Considerations 6-2
6.1.1.1. Epidemiologic Studies 6-2
6.1.1.2. Experimental Studies 6-4
6.2. Cardiovascular and Systemic Effects 6-8
6.2.1. Heart Rate and Heart Rate Variability 6-8
6.2.1.1. Epidemiologic Studies 6-9
6.2.1.2. Human Clinical Studies 6-16
6.2.1.3. Toxicological Studies 6-19
6.2.2. Arrhythmia 6-22
6.2.2.1. Epidemiologic Studies 6-22
6.2.2.2. Toxicological Studies 6-29
6.2.3. Ischemia 6-31
6.2.3.1. Epidemiologic Studies 6-32
6.2.3.2. Human Clinical Studies 6-33
6.2.3.3. Toxicological Studies 6-33
6.2.4. Vasomotor Function 6-35
6.2.4.1. Epidemiologic Studies 6-36
6.2.4.2. Human Clinical Studies 6-38
6.2.4.3. Toxicological Studies 6-41
6.2.5. Blood Pressure 6-45
6.2.5.1. Epidemiologic Studies 6-46
6.2.5.2. Human Clinical Studies 6-49
6.2.5.3. Toxicological Studies 6-50
6.2.6. Cardiac Contractility 6-51
6.2.6.1. Toxicological Studies 6-51
6.2.7. Systemic Inflammation 6-52
6.2.7.1. Epidemiologic Studies 6-52
6.2.7.2. Human Clinical Studies 6-56
6.2.7.3. Toxicological Studies 6-58
6.2.8. Blood Coagulation 6-59
6.2.8.1. Epidemiologic Studies 6-60
6.2.8.2. Human Clinical Studies 6-61
6.2.8.3. Toxicological Studies 6-63
6.2.9. Systemic and Cardiac Oxidative Stress 6-65
6.2.9.1. Epidemiologic Studies 6-65
6.2.9.2. Human Clinical Studies 6-66
6.2.9.3. Toxicological Studies 6-67
6.2.10. Hospital Admissions and ED Visits 6-69
6.2.10.1. All Cardiovascular Disease 6-75
6.2.10.2. Cardiac Diseases 6-81
6.2.10.3. Ischemic Heart Disease 6-81
6.2.10.4. Acute Myocardial Infarction 6-84
6.2.10.5. Congestive Heart Failure 6-85
6.2.10.6. Cardiac Arrhythmias 6-87
6.2.10.7. Cerebrovascular Disease 6-88
6.2.10.8. Ischemic Strokes and Transient Ischemic Attacks 6-89
December 2008
vi
DRAFT-DO NOT QUOTE OR CITE
-------
6.2.11. Summary of Causal Determinations by PM Metric 6-94
6.2.11.1. PM10 6-94
6.2.11.2. PM10-2.5 6-97
6.2.11.3. PM2.5 6-98
6.2.11.4. Ultrafine PM 6-105
6.3. Respiratory Effects 6-109
6.3.1. Respiratory Symptoms and Medication Use 6-109
6.3.1.1. Epidemiologic Studies 6-109
6.3.1.2. Human Clinical Studies 6-120
6.3.2. Pulmonary Function 6-121
6.3.2.1. Epidemiologic Studies 6-121
6.3.2.2. Human Clinical Studies 6-125
6.3.2.3. Toxicological Studies 6-126
6.3.3. Pulmonary Inflammation 6-128
6.3.3.1. Epidemiologic Studies 6-128
6.3.3.2. Human Clinical Studies 6-131
6.3.3.3. Toxicological Studies 6-134
6.3.4. Oxidative Responses 6-140
6.3.4.1. Human Clinical Studies 6-141
6.3.4.2. Toxicological Studies 6-141
6.3.5. Pulmonary Injury 6-143
6.3.5.1. Epidemiologic Studies 6-143
6.3.5.2. Toxicological Studies 6-144
6.3.6. Allergic Responses 6-152
6.3.6.1. Human Clinical Studies 6-152
6.3.6.2. Toxicological Studies 6-153
6.3.7. Host Defense 6-159
6.3.7.1. Toxicological Studies 6-159
6.3.8. Respiratory ED Visits, Hospital Admissions and Physician Visits 6-162
6.3.8.1. All Respiratory Diseases 6-164
6.3.8.2. Asthma 6-172
6.3.8.3. COPD 6-177
6.3.8.4. Pneumonia and Respiratory Infections 6-179
6.3.8.5. Copollutant Models 6-182
6.3.9. Summary and Causal Determinations 6-183
6.3.9.1. PM10 6-183
6.3.9.2. PM10-2.5 6-184
6.3.9.3. PM2.5 6-186
6.3.9.4. Ultrafine Particles 6-193
6.4. Central Nervous System Effects 6-195
6.4.1. Human Clinical Studies 6-195
6.4.2. Toxicological Studies 6-196
6.4.3. Summary and Causal Determination 6-198
6.5. Mortality Associated with Short-Term Exposure 6-198
6.5.1. Summary of Findings from 2004 PM 6-199
6.5.2. Associations of Mortality and Short-Term Exposure to PM 6-201
6.5.2.1. PM10 6-202
6.5.2.2. PM2.5 6-218
6.5.2.3. Other Size-fractionated PM Indices 6-225
6.5.2.4. Ultrafine Particles 6-228
6.5.2.5. Chemical Components of PM 6-229
6.5.2.6. Use of Source-Apportioned PM 6-236
6.5.2.7. Investigation of Concentration-Response Relationship 6-237
6.5.3. Summary of Causal Determinations by PM Metric 6-240
6.5.3.1. PM10 6-240
6.5.3.2. PM2.5 6-241
6.5.3.3. PM10-2.5 6-242
6.5.3.4. Ultra-fine particles (UFP: diameter: 0.01-0.1 |jm) 6-242
6.6. Attribution of Health Effects to Specific Constituents or Sources 6-242
6.6.1. Evaluation Approach 6-243
December 2008
vii
DRAFT-DO NOT QUOTE OR CITE
-------
6.6.2. Findings 6-245
6.6.2.1. Results from the 2004 PM AQCD 6-245
6.6.2.2. Epidemiologic Studies 6-245
6.6.2.3. Human Clinical Studies 6-250
6.6.2.4. Toxicological Studies 6-251
6.6.3. Summary 6-255
Chapter 7. Integrated Health Effects of Long-Term PM Exposure 7-1
7.1. Introduction 7-1
7.2. Cardiovascular and Systemic Effects 7-1
7.2.1. Atherosclerosis 7-2
7.2.1.1. Epidemiologic Studies 7-2
7.2.1.2. Toxicological Studies 7-5
7.2.2. Thromboembolism 7-9
7.2.2.1. Epidemiologic Studies 7-9
7.2.3. Systemic Inflammation and Blood Coagulation 7-9
7.2.3.1. Toxicological Studies 7-9
7.2.4. Renal and Vascular Function 7-10
7.2.4.1. Epidemiologic Studies 7-11
7.2.4.2. Toxicological Studies 7-13
7.2.5. Autonomic Function 7-14
7.2.6. Clinical Outcomes in Epidemiologic Studies 7-14
7.2.7. Overall Summary and Causal Determination 7-18
7.2.7.1. PM10 7-18
7.2.7.2. PMi0-2.5 7-20
7.2.7.3. PM2.5 7-20
7.2.7.4. Ultrafine PM 7-22
7.3. Respiratory Effects 7-23
7.3.1. Respiratory Symptoms and Disease Incidence 7-24
7.3.1.1. Epidemiologic Studies 7-24
7.3.2. Pulmonary Function 7-31
7.3.2.1. Epidemiologic Studies 7-31
7.3.2.2. Toxicological Studies 7-37
7.3.3. Pulmonary Inflammation 7-39
7.3.3.1. Epidemiologic Studies 7-39
7.3.3.2. Toxicological Studies 7-39
7.3.4. Pulmonary Oxidative Response 7-42
7.3.4.1. Toxicological Studies 7-42
7.3.5. Pulmonary Injury 7-43
7.3.5.1. Toxicological Studies 7-43
7.3.6. Allergic Responses 7-46
7.3.6.1. Toxicological Studies 7-46
7.3.7. Host Defense 7-47
7.3.7.1. Toxicological Studies 7-47
7.3.8. Summary and Causal Determination 7-48
7.3.8.1. PM10 7-48
7.3.8.2. PM2.5 7-49
7.3.8.3. Ultrafine PM 7-52
7.4. Reproductive, Developmental, Prenatal and Neonatal Outcomes 7-52
7.4.1. Epidemiologic Studies 7-52
7.4.2. Toxicological Studies 7-73
7.4.3. Summary and Causal Determination 7-82
7.4.3.1. PM10 7-82
7.4.3.2. PM2.5 7-84
7.4.3.3. PM10-2.5 7-86
7.5. Cancer Incidence, Mutagenicity, and Genotoxicity 7-86
7.5.1. Epidemiologic Studies 7-86
7.5.1.1. Toxicological Studies 7-90
7.5.2. Summary and Causal Determinations 7-100
December 2008
viii
DRAFT-DO NOT QUOTE OR CITE
-------
7.6. Mortality Associated with Long-term Exposure 7-100
7.6.1. Review of 1996 and 2004 PM AQCDs 7-100
7.6.2. PM2.5 7-102
7.6.3. PM10-2.5 7-108
7.6.4. PM10 7-109
7.6.5. Composition and Source-Oriented Analyses of PM 7-110
7.6.6. Within-City Effects of PM Exposure 7-111
7.6.7. Effects of Different Long-term Exposure Windows 7-113
7.6.8. Summary and Causal Determinations 7-115
Chapter 8. Public Health Impacts 8-1
8.1. Concentration-Response Relationship 8-1
8.1.1. Mortality Associated with Short-Term Exposure to PM 8-2
8.1.2. Mortality Associated with Long-Term Exposure to PM 8-2
8.1.3. Summary of Concentration-Response Relationship 8-3
8.2. Potentially Susceptible and Vulnerable Subpopulations 8-3
8.2.1. Susceptibility Characteristics 8-5
8.2.1.1. Age 8-5
8.2.1.2. Pregnancy 8-7
8.2.1.3. Gender 8-8
8.2.1.4. Race/Ethnicity 8-8
8.2.1.5. Gene-Environment Interaction 8-9
8.2.1.6. Pre-Existing Disease 8-10
8.2.1.7. Cardiovascular Diseases 8-11
8.2.1.8. Respiratory Illnesses 8-13
8.2.1.9. Respiratory Contributions to CV Effects 8-14
8.2.1.10. Inflammatory Conditions: Diabetes and Obesity 8-15
8.2.2. Vulnerability Characteristics 8-17
8.2.3. Urban Environment 8-17
8.2.4. Socioeconomic Status 8-17
8.2.5. Geographic Location 8-19
Chapter 9. Ecosystem and Welfare Effects 9-1
9.1. Introduction 9-1
9.2. Summary and Conclusions 9-1
9.2.1. Summary of Effects on Visibility 9-1
9.2.2. Summary of Effects on Individual Organisms and Ecosystems 9-5
9.2.3. Summary of Effects on Materials 9-6
9.2.4. Summary of Effects on Climate 9-7
9.3. Effects on Visibility 9-7
9.3.1. Introduction 9-7
9.3.2. Background 9-8
9.3.2.1. Non-PM Visibility Effects 9-12
9.3.2.2. PM Visibility Effects 9-13
9.3.3. Effects on Visibility 9-16
9.3.4. Monitoring and Assessment 9-16
9.3.4.1. Aerosol Properties 9-16
9.3.4.2. Spatial Patterns 9-23
9.3.4.3. Urban and Regional Patterns 9-31
9.3.4.4. Temporal Trends 9-40
9.3.4.5. Causes of Haze 9-45
9.3.5. Urban Visibility Valuation and Preference 9-74
9.3.5.1. Urban Visibility Preference Studies 9-76
9.3.5.2. Denver, Colorado Urban Visibility Preference Study 9-78
9.3.5.3. Phoenix, Arizona Urban Visibility Preference Study 9-79
9.3.5.4. British Columbia, Canada Urban Visibility Preference Study 9-79
9.3.5.5. Washington, DC Urban Visibility Pilot Preference Study 9-80
9.3.5.6. Urban Visibility Valuation Studies 9-82
December 2008
ix
DRAFT-DO NOT QUOTE OR CITE
-------
9.4. Deposition of PM 9-83
9.4.1. Forms of Deposition 9-84
9.4.1.1. Fine vs. Coarse PM 9-84
9.4.1.2. Deposition Modes 9-85
9.4.2. Methods for Estimating Dry Deposition 9-87
9.4.3. Factors Affecting Dry Deposition 9-90
9.4.3.1. Leaf Surface Effects on Deposition Velocity 9-92
9.4.3.2. Canopy Surface Effects on Deposition Velocity 9-92
9.4.4. Magnitude of Dry Deposition 9-93
9.4.4.1. Using Vegetation for Estimating Atmospheric Deposition 9-94
9.4.4.2. Deposition to Canopies 9-96
9.4.4.3. Deposition to Soil 9-97
9.4.5. Components of Deposition 9-98
9.4.5.1. Trace Metals 9-98
9.4.5.2. Mercury 9-101
9.4.5.3. Organics 9-103
9.4.5.4. Base Cations 9-105
9.5. Effects on Individual Organisms 9-106
9.5.1. Effects on Plants 9-106
9.5.1.1. Direct Effects of Coarse-mode Particles 9-108
9.5.1.2. Effects of Fine-mode Particles 9-109
9.5.2. Effects on Animals 9-116
9.5.3. Effects on Microbes and Fungi 9-117
9.6. Effects on Ecosystems 9-119
9.6.1. Biogeochemical Processes 9-121
9.6.2. Bioaccumulation 9-122
9.6.2.1. Metals 9-122
9.6.2.2. Organics 9-124
9.6.3. Nutrient Cycling 9-125
9.6.4. Ecosystem Structure and Function 9-126
9.7. Effects on Materials 9-127
9.7.1. Effects on Paint 9-130
9.7.2. Effects on Metal Surfaces 9-130
9.7.3. Effects on Stone 9-131
9.8. Effects on Climate 9-132
9.8.1. Direct Effects 9-136
9.8.1.1. Radiation Budget 9-136
9.8.1.2. Temperature 9-141
9.8.1.3. Precipitation 9-143
9.8.1.4. Magnitude of Overall Direct Effects 9-143
9.8.2. Indirect Effects 9-145
9.8.2.1. First Indirect Effect: Cloud Albedo 9-146
9.8.2.2. Second Indirect Effect: Cloud Lifetime 9-150
9.8.3. Other Effects 9-152
9.8.4. Effects on Local and Regional Climate 9-152
9.8.5. Glaciers and Snowpack 9-154
9.8.6. Global Warming Potentials 9-157
References R-1
December 2008
x
DRAFT-DO NOT QUOTE OR CITE
-------
List of Tables
Table 1-1. Summary of NAAQS promulgated for PM, 1971-2006. 1-5
Table 1-2. Aspects to aid in judging causality. 1-19
Table 1-3. Weight of evidence for causal determination. 1-22
Table 2-1. Study-specific PM2.5 factor/source categories associated with health effects. 2-21
Table 3-1. Characteristics of ambient fine (ultrafine plus accumulation-mode) and coarse particles. 3-6
Table 3-2. Constituents of atmospheric particles and their major sources. 3-9
Table 3-3. Proximity to PM2.5 monitors for the total population by city. 3-32
Table 3-4. Proximity to PM10 monitors for the total population by city. 3-32
Table 3-5. Proximity to PM2.5 monitors for children aged 0-4 by city. 3-33
Table 3-6. Proximity to PM10 monitors for children aged 0-4 by city. 3-34
Table 3-7. Proximity to PM2.5 monitors for children aged 5-17 by city. 3-34
Table 3-8. Proximity to PM10 monitors for children aged 5-17 by city. 3-35
Table 3-9. Proximity to PM2.5 monitors for adults aged 65 and older by city. 3-36
Table 3-10. Proximity to PM10 monitors for adults aged 65 and older by city. 3-36
Table 3-11. PM10 distributions derived from AQS data (concentration in |jg/m3). 3-40
Table 3-12. PM2.5 distributions derived from AQS data (concentration in |jg/m3). 3-44
Table 3-13. PM10-2.5 distributions derived from AQS data (concentration in |jg/m3). 3-46
Table 3-14. Inter-sampler correlation statistics for each pair of PM10 AQS data for Boston, MA. 3-60
Table 3-15. Inter-sampler correlation statistics for each pair of PM10 AQS data for Pittsburgh, PA. 3-63
Table 3-16. Inter-sampler correlation statistics for each pair of PM10 AQS data for Los Angeles, CA. 3-65
Table 3-17. Inter-sampler correlation statistics for each pair of PM2.5 AQS data for Boston, MA. 3-71
Table 3-18. Inter-sampler correlation statistics for each pair of PM2.5 AQS data for Pittsburgh, PA. 3-74
Table 3-19. Inter-sampler correlation statistics for each pair of PM2.5 AQS data for Los Angeles, CA. 3-76
Table 3-20. Emissions factors (ng/kg) for trace elements under variable speed and steady speed driving conditions for PM
emitted by diesel and gasoline engines. 3-106
Table 3-21. Estimates of annual average natural background concentrations of PM in different size fractions (|jg/m3) from
previous reviews. 3-117
Table 3-22. Annual and quarterly mean PM2.5 concentrations (|jg/m3) measured at IMPROVE sites in 2004. 3-118
Table 3-23. Annual and quarterly mean PM2.5 concentrations (|jg/m3) for the CMAQ "base case" at IMPROVE sites in 2004. 3-127
Table 3-24. Annual and quarterly mean PM2.5 concentrations (|jg/m3) for the CMAQ PRB simulations at IMPROVE sites in
2004. 3-127
December 2008
xii
DRAFT-DO NOT QUOTE OR CITE
-------
Table 3-25. Annual and quarterly mean of the CMAQ-predicted base case PM2.5 concentrations (|jg/m3) in the U.S. EPA
CONUS regions in 2004. 3-128
Table 3-26. Annual and quarterly mean of the CMAQ-predicted PRB PM2.5 concentrations (|jg/m3) in the U.S. EPA CONUS
regions in 2004. 3-128
Table 3-27. Statistical parameters for the relationships between exposures and ambient concentrations for each subject
separately (values of C based on average of five monitoring sites, E and A outliers included) sorted according to
the correlation coefficient of E with C. 3-132
Table 3-28. Examples of studies comparing outdoor personal exposures with fixed site ambient concentrations. 3-144
Table 3-29. Proximity to PM10 and PM2.5 monitors among adults older than 25 with less than a high school education by city.
Percentages are given with respect to the total population per city provided. 3-162
Table 3-30. Proximity to PM10 and PM2.5 monitors for the total population under poverty line by city. Percentages are given with
respect to the total population per city provided. 3-163
Table 3-31. Proximity to PM10 and PM2.5 monitors for adults older than 25 with at least a high school degree by city.
Percentages are given with respect to the total population per city provided. 3-164
Table 3-32. Proximity to PM10 and PM2.5 monitors for adults older than 25 with at least a college degree by city. Percentages
are given with respect to the total population per city provided. 3-165
Table 6-1. Characteristics of epidemiologic/panel studies investigating associations between PM and changes in HRV. 6-14
Table 6-2. Studies of ventricular arrhythmia and ambient PM concentration, in patients with implantable cardioverter
defibrillators. 6-25
Table 6-3. Median particulate concentration. 6-46
Table 6-4. Ambient concentrations in six European cities. 6-53
Table 6-5. Description of ICD-9 and ICD-10 codes for diseases of the circulatory system. 6-71
Table 6-6. Characterization of ambient PM concentrations in studies of hospital admission and ED visits for cardiovascular
diseases. 6-80
Table 6-7. Characterization of ambient PM concentrations from studies of respiratory outcomes and short-term exposures in
asthmatic adults. 6-114
Table 6-8. PAMCHAR PM10-2.5 inflammation results with ambient PM. 6-149
Table 6-9. Other ambient PM - in vivo PM10-2.5 studies - BALF results, 18-24 h post-IT 6-150
Table 6-10. Description of ICD-9 and ICD-10 codes for diseases of the respiratory system. 6-163
Table 6-11. PM concentrations in studies of respiratory diseases published since 2002. 6-164
Table 6-12. Characterization of ambient PM concentrations from studies of hospitalization or ED visits for respiratory diseases 6-182
Table 6-13. Overview of U.S. and Canadian multicity PM studies analyzed in the 2004 PM AQCD and the PM ISAb 6-200
Table 6-14. NMMAPS national and regional percentage increase in all-cause, cardio-respiratory, and other-cause mortality
associated with a 10 |jg/m3 increase in PM10 at lag 1 day for the periods 1987-1994,1995-2000, and 1987-2000. 6-205
Table 6-15. Effect modification of composition on the estimated percent increase in mortality with a 10 |jg/m3 increase in PM2.5. 6-233
Table 6-16. Epidemiologic studies of PM sources, factors, or individual components. 6-247
Table 6-17. Human clinical studies of PM sources, factors, or individual components. 6-250
Table 6-18. Toxicological studies of PM sources, factors, or individual constituents 6-252
Table 7-1. Characterization of ambient PM concentrations from studies of subclinical measures of cardiovascular diseases. 7-13
December 2008
xiii
DRAFT-DO NOT QUOTE OR CITE
-------
Table 7-2. Characterization of ambient PM concentrations from studies of clinical cardiovascular diseases. 7-23
Table 7-3. Characterization of ambient PM concentrations from studies of respiratory symptoms/disease and long-term
exposures. 7-26
Table 7-4. Characterization of ambient PM concentrations from studies of FEVi and long-term exposures. 7-32
Table 7-5. Characterization of ambient PM concentrations from studies of reproductive, developmental, prenatal and neonatal
outcomes and long-term exposure. 7-54
Table 7-6. Characterization of ambient PM concentrations from select studies of cancer and long-term exposures. 7-87
Table 7-7. Association of average air pollution concentrations and traffic variables with lung cancer incidence in full cohort
and case-cohort analyses. 7-87
Table 7-8. Characterization of ambient PM concentrations from studies of mortality and long-term exposures. 7-102
Table 7-9. Distribution of the effect of a hypothetical reduction of 10 |jg/m3 PM10 in 2000 on all-cause mortality 2000-2009 in
Switzerland. 7-115
Table 8-1. Characteristics of susceptible/vulnerable subpopulations. 8-4
Table 8-2. Percent of the U.S. population inflicted with respiratory diseases, cardiovascular diseases, and diabetes. 8-16
Table 9-1. Regional Planning Organization websites with visibility characterization and source attribution assessment
information. 9-24
Table 9-2. Summary of urban visibility preference studies. 9-77
Table 9-3. Factors potentially important in estimating mercury exposure and how they are addressed in this study. 9-89
Table 9-4. Range in estimated source strength (Tg aerosol/year). 9-138
Table 9-5. Overview of the different aerosol indirect effects and their sign of the net radiative flux change at the top of the
atmosphere (TOA). 9-145
Table 9-6. Overview of the different aerosol indirect effects and their implications for the global mean net shortwave radiation
of the surface Fsfc (columns 2-4) and for precipitation (columns 5-7). 9-146
Table 9-7. Recent studies highlighting POP occurrence and fate in the major arctic compartments. 9-156
December 2008
xiv
DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
Figure 3-1. Particle size distributions by number and volume. 3-5
Figure 3-2. Detailed source categorization of emissions of primary PM2.5, PM10 and gaseous precursor species SO2, NOx, NH3
and VOCs for 2002 in units of million metric tons. 3-11
Figure 3-3. Primary emissions and formation of secondary organic aerosol through gas, cloud and condensed phase
reactions. 3-14
Figure 3-4. PM10 monitor distribution in comparison with population density, Boston CSA. 3-30
Figure 3-5. PM2.5 monitor distribution in comparison with population density, Boston CSA. 3-31
Figure 3-6. Average 24-h PM10 concentration by county derived from FRM or FEM monitors, 2005-2007. 3-40
Figure 3-7. Average 24-h PM2.5 concentration by county derived from FRM or FRM-like data, 2005-2007. 3-43
Figure 3-8. Average 24-h PM10-2.5 concentration by county derived from co-located low volume FRM PM10 and PM2.5 monitors,
2005-2007. 3-45
Figure 3-9. OC concentrations measured at CSN sites across the U.S., 2005-2007. 3-47
Figure 3-10. EC concentrations measured at CSN sites across the U.S., 2005-2007. 3-48
Figure 3-11. SO42" concentrations measured at CSN sites across the U.S., 2005-2007. 3-49
Figure 3-12. NO3" concentrations measured at CSN sites across the U.S., 2005-2007. 3-50
Figure 3-13. NH4"* concentrations measured at CSN sites across the U.S., 2005-2007. 3-51
Figure 3-14. Annual average FRM PM2.5 speciation data for 2005-2007 derived using the SANDWICH method in fifteen
CSAs/CBSAs: 3-54
Figure 3-15. Seasonally averaged FRM PM2.5 speciation data for 2005-2007 for winter derived using the SANDWICH method in
fifteen CSAs/CBSAs: 3-55
Figure 3-16. Seasonally averaged FRM PM2.5 speciation data for 2005-2007 for spring derived using the SANDWICH method
in fifteen CSAs/CBSAs: 3-56
Figure 3-17. Seasonally averaged FRM PM2.5 speciation data for 2005-2007 for summer derived using the SANDWICH method
in fifteen CSAs/CBSAs: 3-56
Figure 3-18. Seasonally averaged FRM PM2.5 speciation data for 2005-2007 for fall derived using the SANDWICH method in
fifteen CSAs/CBSAs: 3-57
Figure 3-19. Map of PM10 FRM distribution with AQS Site IDs for Boston, MA. 3-59
Figure 3-20. Box plot illustrating the seasonal distribution of 24-h average PM10 concentrations for Boston, MA. 3-60
Figure 3-21. Map of PM10 FRM distribution with AQS Site IDs for Pittsburgh, PA. 3-61
Figure 3-22. Box plots illustrating the seasonal distribution of 24-h average PM10 concentrations in Pittsburgh, PA. 3-62
Figure 3-23. Map of PM10 FRM distribution with AQS Site IDs for Los Angeles, CA. 3-64
Figure 3-25. PM10 inter-sampler correlations as a function of distance between monitors for Boston. 3-67
Figure 3-26. PM10 inter-sampler correlations as a function of distance between monitors for Pittsburgh. 3-67
Figure 3-27. PM10 inter-sampler correlations as a function of distance between monitors for Los Angeles. 3-68
December 2008
xv
DRAFT-DO NOT QUOTE OR CITE
-------
Figure 3-28. PM2.5 monitor distribution in comparison with source distribution, Boston, MA. 3-69
Figure 3-29. Box plots illustrating the seasonal distribution of 24-h average PM2.5 concentrations for Boston, MA. 3-70
Figure 3-30. PM2.5 monitor distribution in comparison with source distribution, Pittsburgh, PA. 3-72
Figure 3-32. PM2.5 monitor distribution in comparison with source distribution, Los Angeles, CA. 3-75
Figure 3-34. PM2.5 inter-sampler correlations as a function of distance between monitors for Boston. 3-78
Figure 3-35. PM2.5 inter-sampler correlations as a function of distance between monitors for Pittsburgh. 3-79
Figure 3-36. PM2.5 inter-sampler correlations as a function of distance between monitors for Los Angeles. 3-79
Figure 3-37. PM10-2.5 generated from all available co-located FRM PM10 and PM2.5 monitors in Atlanta, Boston, Chicago,
Denver, New York and Phoenix, 2005-2007. 3-80
Figure 3-38. Bin-wise Spearman correlation coefficients in aerosol particle number concentrations between the Ift (urban
background) and the Eisenbahn-strasse (city/urban center) sites in Leipzig, Germany. 3-82
Figure 3-39. Dimensionless concentration as a function of height at windward and leeward locations and street canyon aspect
ratios (HOT). 3-83
Figure 3-40. Mass distributions for BaP at a high traffic urban center (HTC), high traffic urban periphery (HTP), low traffic urban
center (LTC), low traffic urban periphery (LTP), and low traffic industrial urban periphery (LTIP) in Seville, Spain. 3-84
Figure 3-41. Mass distributions for sixteen PAHs at a high traffic city center in Seville, Spain. 3-85
Figure 3-42. Figure to be replaced. Particle size distributions measured at various distances from the 710 freeway in Los
Angeles (top), and particle number concentration as a function of distance from the 710 freeway (bottom). 3-87
Figure 3-43. Inter-sampler correlations as a function of distance between monitors for samplers located within 4 km
(neighborhood scale) for PM2.5 and PM10. 3-88
Figure 3-44. Ambient 24-h PM10 concentrations in the U.S., 1988-2007, showing (A) ambient concentrations and (B) number of
trends sites above the NAAQS. 3-89
Figure 3-45. Ambient 24-h PM10 concentrations in the contiguous U.S. by EPA region, 1988-2007. 3-90
Figure 3-46. Ambient 24-h PM2.5 concentrations in the U.S., 1999-2007, showing (A) ambient concentrations and (B) number of
trends sites above the NAAQS. 3-91
Figure 3-47. Ambient 24-h PM2.5 concentrations in the contiguous U.S. by EPA region, 1999-2007. 3-92
Figure 3-48. Ambient annual PM2.5 concentrations in the U.S., 1999-2007, showing (A) ambient concentrations and (B) number
of trends sites above the NAAQS. 3-93
Figure 3-49. Ambient annual PM2.5 concentrations in the contiguous U.S. by EPA region, 1999-2007. 3-94
Figure 3-50. Ultrafine particle size distribution at highway (site A) and background (site B) sites in Los Angeles during summer
and winter seasons, with winter broken into day and night distributions. 3-95
Figure 3-51. Diel plot generated from hourly FEM PM10 data (|ig/m3) stratified by weekday (left) and weekend (right) for
Chicago and Phoenix from 2005 to 2007. 3-97
Figure 3-52. Diel plot generated from hourly FRM-like PM2.5 data (|ig/m3) stratified by weekday (left) and weekend (right) for
Pittsburgh and Seattle from 2005 to 2007. 3-98
Figure 3-53. Average diurnal variation of NOx, CO, particle number and particle volume on weekdays (left) and Sundays (right). 3-99
Figure 3-54. Correlations between 24-h PM10 and co-located 24-h average PM2.5, PM10-2.5, SO2, NO2 and CO and daily max 8-h
avg O3 for the U.S. stratified by season (2005-2007). 3-100
Figure 3-55. Correlations between 24-h PM2.5 and co-located 24-h average PM10, PM10-2.5, SO2, NO2 and CO and daily max 8-h
avg O3 for the U.S. stratified by season (2005-2007). 3-101
December 2008
xvi
DRAFT-DO NOT QUOTE OR CITE
-------
Figure 3-56. Schematic of organic composition of particulate emissions from gasoline-fueled vehicles. 3-105
Figure 3-57 Source category contributions to PM2.5 at a number of sites in the East derived using PMF. 3-108
Figure 3-58. Pearson correlation coefficients for source category contributions to PM2.5 at the ten Regional Air Pollution
Study/Regional Air Monitoring System (RAPS/RAMS) monitoring sites in St. Louis. 3-109
Figure 3-59. Pearson correlation coefficients for source contributions to PM10-2.5 at the ten Regional Air Pollution
Study/Regional Air Monitoring System (RAPS/RAMS) monitoring sites in St. Louis. 3-110
Figure 3-60. IMPROVE monitoring site locations. 3-117
Figure 3-61. 12-km EUS Summer SO42" PM each data point represents a paired monthly averaged (June/July/August)
observation and CMAQ prediction at a particular IMPROVE, STN, and CASTNet site. 3-120
Figure 3-62 12-km EUS Winter nitrate PM, each data point represents a paired monthly averaged
(December/January/February) observation and CMAQ prediction at a particular IMPROVE and STN site. 3-121
Figure 3-63. 12-km EUS Winter total nitrate (HNO3 + total PNO3), each data point represents a paired monthly averaged
(December/January/February) observation and CMAQ prediction at a particular CASTNet site. 3-121
Figure 3-64. Monthly average of PM2.5 concentrations measured at IMPROVE sites in the East and Midwest for 2004. 3-122
Figure 3-65. Monthly average of PM2.5 concentrations measured at IMPROVE sites in the West for 2004. 3-123
Figure 3-66. Monthly average of PM2.5 concentrations measured at the Redwoods National Park IMPROVE sites in California
for 2004. 3-123
Figure 3-67. Distribution of PM2.5 concentrations measured at IMPROVE sites in the East and Midwest for 2004. Also shown
are distributions of PM2.5 concentrations calculated by CMAQ for the base case and for PRB. 3-124
Figure 3-68. Distribution of PM2.5 concentrations measured at IMPROVE sites in the West for 2004. Also shown are
distributions of PM2.5 concentrations calculated by CMAQ for the base case and for PRB. Note the scale change
from the preceding figures. 3-125
Figure 3-69. Distribution of PM2.5 concentrations measured at the Redwoods National Park IMPROVE sites in California for
2004. Also shown are distributions of PM2.5 concentrations calculated by CMAQ for the base case and for PRB.
Note the scale change from the preceding figures. 3-126
Figure 3-70. Distribution of time sample population spends in various environments, from the National Human Activity Pattern
Survey. 3-130
Figure 3-71. Comparison of community dichot and personal exposure monitor for PM10-2.5. 3-137
Figure 3-72. Grid resolution of the CMAQ model in Philadelphia compared with distribution of census tracts in which exposure
assessment is performed. 3-142
Figure 3-73. Time series plot of PM2.5 concentration measured at various sites during September - November 2001. 3-146
Figure 3-74. Total exposure to SO4 as a function of measured ambient SO4 concentration. 3-149
Figure 3-75. Estimated ambient exposure to PM2.5 as a function of measured ambient PM2.5 concentration. 3-149
Figure 3-76. Total exposure to PM2.5 as a function of measured ambient PM2.5 concentration. 3-150
Figure 3-77. Apportionment of aliphatic carbon, carbonyl, and SO4"2 components of outdoor, indoor, and personal PM2.5
samples, for Los Angeles (top), Elizabeth (center), and Houston (bottom). 3-153
Figure 3-78. Apportionment of infiltrated mechanically-generated (top), primary combustion (center), and secondary combustion
(bottom). 3-154
Figure 3-79. Fw as a function of particle size. 3-157
Figure 3-80. Results of the positive matrix factorization model showing differences in the composition of outdoor PM and PM
that has infiltrated indoors. 3-158
December 2008
xvii
DRAFT-DO NOT QUOTE OR CITE
-------
Figure 3-81. PM2.5 sampler density compared with numbers in poverty per square mile in the Philadelphia CSA. 3-166
Figure 4-1. Diagrammatic representation of respiratory tract regions in humans. 4-3
Figure 4-2. Structure of lower airways with progression from the large airways to the alveolus. 4-4
Figure 4-3 Comparison of total and regional deposition results from the ICRP and the MPPD models for a resting breathing
pattern (Vt = 625 ml, f= 12 min-1). 4-8
Figure 4-4 Comparison of total and regional deposition results from the ICRP and the MPPD models for a light exercise
breathing pattern (Vt = 1250 ml, f= 20 min-1). 4-9
Figure 4-5. Total lung deposition measured in healthy adults (ultrafine, 11 M, 11 F, 31 ±4 years; fine and coarse, 11 M, 11 F,
25 ± 4 years) during controlled breathing on a mouthpiece. 4-10
Figure 4-6. Total deposition of hygroscopic sodium chloride and hydrophobic aluminosilicate aerosols during oral breathing
(Vt= 1.0 L, f = 15 min-1). 4-18
Figure 4-7. Retention of poorly soluble particles (0.5-5 |jm) in the alveolar region of the lung over various mammalian species. 4-23
Figure 5-1. PM oxidative potential. 5-2
Figure 5-2. PM stimulates pulmonary cells to produce ROS/RNS. 5-3
Figure 5-3. PM activates cell signaling pathways leading to pulmonary inflammation. 5-4
Figure 5-4. Potential pathways for the effects of PM on the respiratory system. 5-7
Figure 5-5. Potential pathways for the effects of PM on the cardiovascular system. 5-10
Figure 6-1. Excess risk estimates per 10 |jg/m3 increase in PM10, PM2.5 and PM10 2.5 for studies of CVD ED visits* and
hospitalizations. 6-79
Figure 6-2. Excess risk estimates per 10 |jg/m3 increase in PM10, PM2.5, PM10-2.5 for studies of ED visits * and hospitalizations
for IHD and Ml. 6-83
Figure 6-3. Excess risk estimates per 10 |jg/m3 increase in PM10, PM10-2.5 and PM10-2.5 for studies of CHF ED visits* and
hospitalizations. 6-87
Figure 6-4. Excess risks estimates per 10 |jg/m3 increase in PM10, PM2.5, and PM10-2.5 for studies of ED visits* and
hospitalizations for cerebrovascular diseases. 6-91
Figure 6-5. Respiratory symptoms and/or medication use among asthmatic children following acute exposure to PM10. 6-110
Figure 6-6. Respiratory symptoms and/or medication use among asthmatic children following acute exposure to PM2.5. 6-111
Figure 6-7. Respiratory symptoms and/or medication use among asthmatic adults following acute exposure to particles. 6-118
Figure 6-8 Respiratory symptoms following acute exposure to particles and additional criteria pollutants. 6-119
Figure 6-9. Excess risks estimates per 10 |jg/m3 24-h average PM10 concentration for studies of ED visits and hospitalizations
for respiratory diseases. 6-167
Figure 6-10. Excess risks estimates per 10 |jg/m3 increase in 24-h average PM2.5 and PM10-2.5 for studies of ED visits and
hospitalizations for respiratory diseases. 6-170
Figure 6-11. Excess risks estimates per 10 |jg/m3 increase in 24-h average PM10, PM2.5 and PM10-2.5 for studies of asthma ED
visits* and hospitalizations. 6-176
Figure 6-12. Excess risks estimates per 10 |jg/m3 increase in 24-h average PM10, PM2.5 and PM10-2.5 for studies of COPD ED
visits* and hospitalizations among older adults (65+years, unless other age group is noted). 6-178
Figure 6-13. Excess risks estimates per 10 |jg/m3 increase in 24-h average PM10, PM2.5 and PM10-2.5 for studies of respiratory
infection ED visits* and hospitalizations. 6-181
December 2008
xviii
DRAFT-DO NOT QUOTE OR CITE
-------
Figure 6-14. National and regional estimates of smooth seasonal effects for PM10 at a 1-day lag and their sensitivity to the
degrees of freedom assigned to the smooth function of time in the updated NMMAPS data 1987-2000. 6-204
Figure 6-15. Effect modification by city characteristics in 20 U.S. cities. 6-210
Figure 6-16. PM10 risk estimates (per 10 |jg/m3) by individual-level characteristics. 6-212
Figure 6-17. PM10 risk estimates (per 10 |jg/m3) by location of death and by season. 6-213
Figure 6-18. PM10 risk estimates (per 10 |jg/m3) by contributing causes of deaths. 6-214
Figure 6-19. Summary of PM10 risk estimates (per 10 |jg/m3) for all-cause mortality from recent multicity studies. 6-216
Figure 6-20. Summary of PM10 risk estimates (per 10 |jg/m3) for cause-specific mortality for all U.S.- and Canadian-based
studies. 6-217
Figure 6-21. Summary of all-cause mortality PM2.5 risk estimates per 10 |jg/m3 by various effect modifiers. 6-221
Figure 6-22. Summary of PM2.5 risk estimates per 10 |jg/m3 for major underlying causes of death. 6-223
Figure 6-23. Summary of PM2.5 risk estimates (per 10 |jg/m3) for cause-specific mortality for all U.S.- and Canadian-based
studies. 6-224
Figure 6-24. Summary of PM10-2.5 risk estimates (per 10 |jg/m3) for cause-specific mortality for all U.S.-, Canadian-, and
international-based studies. 6-228
Figure 6-25. Percent increase in PM10 risk estimates (point estimates and 95% confidence intervals) associated with a
5th-to-95th percentile: increase in PM2.5 and PM2.5 chemical components. 6-230
Figure 6-26. Sensitivity of the percent increase in PM10 risk estimates (point estimates and 95% confidence intervals)
associated with an interquartile increase in Ni. 6-231
Figure 6-27. Excess risk (CI) of total mortality per IQR of concentrations. 6-235
Figure 6-28. Relative risk and CI of cardiovascular mortality associated with estimated PM2.5 source contributions. 6-237
Figure 6-29. Concentration-response curves (spline model) for all-cause, cardiovascularrespiratory and other cause mortality
from the 20 NMMAPS cities. 6-238
Figure 6-30. Percent increase in the risk death on days with PM10 concentrations in the ranges of 15-24, 25-34,35-44, and
45 |jg/m3 and greater, compared to a reference of days when concentrations were below 15 |jg/m3. 6-239
Figure 6-31. Combined concentration-response curves (spline model) for all-cause, cardiovascular, and respiratory mortality
from the 22 APHEA cities. 6-240
Figure 7-1. Risk estimates for the associations of clinical outcomes with long-term exposure to ambient PM2.5 and PM10 7-16
Figure 7-2. Adjusted ORs and 95% CIs of symptoms and respiratory diseases in Swiss Surveillance Program of Childhood
Allergy and Respiratory Symptoms with respect to air pollution and climate associated with a decline of 10 ug/m3
PM10 levels (A)1. 7-27
Figure 7-3. Effect of PM2.5 on the association of lung function with asthma. 7-30
Figure 7-4. Decrements in FEV1, FVC, FEFso%, FEF25-75, and MMEF and a 10 |jg/m3 change in PM10. 7-32
Figure 7-5. Proportion of 18-year olds with a FEV1 below 80% of the predicted value plotted against the average levels of
pollutants from 1994 through 2000 in the 12 southern California communities of the Children's Health Study. 7-35
Figure 7-6. Percent increase in postneonatal mortality per 10 |jg/m3 in PM10, comparing risk for total and respiratory mortality. 7-72
Figure 7-7. Mortality risk estimates associated with long-term exposure to PM2.5 from the Harvard Six Cities Study (SCS) and
the American Cancer Society Study (ACS). 7-103
Figure 7-8. Mortality risk estimates associated with long-term exposure to PM2.5 in cohort studies. 7-104
December 2008
xix
DRAFT-DO NOT QUOTE OR CITE
-------
Figure 7-9. Plots of the relative risk of death from cardiovascular disease from the Women's Health Initiative study displaying
the between-city and within-city contributions to the overall association between PM2.5 and cardiovascular mortality
windows of exposure-effects. 7-112
Figure 7-10. The model-averaged estimated effect of a 10- |jg/m3 increase in PM2.5 on all-cause mortality at different lags (in
years) between exposure and death. Each lag is estimated independently of the others. Also shown are the
pointwise 95%Cls for each lag, based on jacknife estimates. 7-113
Figure 7-11. Time course of relative risk of death after a sudden decrease in air pollution exposure during the year 2000,
assuming a steady state model (solid line) and a dynamic model (bold dashed line). The thin dashed line refers to
the reference scenario. 7-114
Figure 7-12. Experts' mean effect estimates and uncertainty distributions for the PM2.5 mortality concentration-response
coefficient for a 1 |jg/m3 change in annual average PM2.5 7-116
Figure 9-1. Important factors involved in seeing a scenic vista are outlined. Image-forming information from an object is
reduced (scattered and absorbed) as it passes through the atmosphere to the human observer. 9-9
Figure 9-2. Schematic of remote-area (top) and urban (bottom) nighttime sky visibility showing the effects of PM and light
pollution. 9-11
Figure 9-3. Effect of relative humidity on light scattering by mixtures of ammonium nitrate and ammonium sulfate. 9-14
Figure 9-4. Estimated fractions of total particulate nitrate during each field campaign comprised of ammonium nitrate, reacted
sea salt nitrate (shown as NaNOs), and reacted soil dust nitrate (shown as Ca(N03)2). 9-19
Figure 9-5. A scatter plot of the original IMPROVE algorithm estimated particle light scattering versus measured particle light
scattering. 9-23
Figure 9-6. Scatter plot of the revised algorithm estimates of light scattering versus measured light scattering. 9-23
Figure 9-7. IMPROVE network PM species estimated light extinction for 2000 (left) and for 2004 (right). 9-25
Figure 9-8. Mean estimated light extinction from PM speciation measurements for the first (top left), second (top right), third
(bottom left), and fourth (bottom right) calendar quarters of 2004. 9-26
Figure 9-9. Percent contributions of ammonium nitrate (left column) and ammonium sulfate (right column) to particulate light
extinction for each calendar quarter of 2004 (first through fourth quarter arranged from top to bottom). 9-27
Figure 9-10. Percent contributions of organic mass (left column) and EC (right column) to particulate light extinction for each
calendar quarter of 2004 (first through fourth quarter arranged from top to bottom). 9-29
Figure 9-11. Percent contributions of coarse mass (left column) and fine soil (right column) to particulate light extinction for
each calendar quarter of 2004 (first through fourth quarter arranged from top to bottom). 9-30
Figure 9-12. IMPROVE Mean PM2.5 mass concentration determined by summing the major components for the 2000 through
2004. 9-31
Figure 9-13. IMPROVE and CSN (STN) mean PM2.5 mass concentration determined by summing the major components for
2000 through 2004 9-32
Figure 9-14. IMPROVE mean ammonium nitrate concentrations for 2000 through 2004. 9-32
Figure 9-15. IMPROVE and CSN (STN) mean ammonium nitrate concentrations for 2000 through 2004. 9-33
Figure 9-16. IMPROVE mean ammonium sulfate concentrations for 2000 through 2004. 9-33
Figure 9-17. IMPROVE and CSN (STN) mean ammonium sulfate concentrations for 2000 through 2004. 9-34
Figure 9-18. IMPROVE monitored mean organic mass concentrations for 2000 through 2004. 9-36
Figure 9-19. IMPROVE and CSN (STN) mean organic mass concentrations for 2000 through 2004. 9-36
Figure 9-20. IMPROVE mean EC concentrations for 2000 through 2004. 9-37
December 2008
xx
DRAFT-DO NOT QUOTE OR CITE
-------
Figure 9-21.
Figure 9-22.
Figure 9-23.
Figure 9-24.
Figure 9-25.
Figure 9-26.
Figure 9-27.
Figure 9-28.
Figure 9-29.
Figure 9-30.
Figure 9-31.
Figure 9-32.
Figure 9-33.
Figure 9-34.
Figure 9-35.
Figure 9-36.
Figure 9-37.
Figure 9-38.
Figure 9-39.
Figure 9-40.
Figure 9-41.
Figure 9-42.
IMPROVE and CSN (STN) mean EC concentrations for 2000 through 2004. 9-37
IMPROVE mean fine soil concentrations for 2000 through 2004. 9-38
IMPROVE and CSN (STN) fine soil concentrations, 2000 through 2004. 9-38
Regional and local contributions to annual average PM2.5 by particulate sulfate, nitrate and total carbon (i.e.
organic plus EC) for select urban areas based on paired IMPROVE and CSN monitoring sites. 9-39
IMPROVE mean coarse mass concentrations for 2000 through 2004. 9-40
Ten-year (1995-2004) haze trends for the mean of the 20% best annual haze conditions. 9-41
Ten-year (1995-2004) haze trends for the mean of the 20% worst annual haze conditions. 9-42
Ten-year trends in the 80th percentile particulate sulfate concentration based on IMPROVE and CASTNet
monitoring and net SO2 emissions from the National Emissions Trends (NET) data base by region of the U.S. 9-43
Map of 10-year trends (1994-2003) in haze by particulate nitrate contribution to haze for the worst 20% annual
haze periods. 9-45
Contributions of the Pacific Coast area to the ammonium sulfate (|jg/m3) at 84 remote-area monitoring sites in
western U.S. based on trajectory regression (dots denote locations of the IMPROVE aerosol monitoring sites). 9-48
Shows the IMPROVE monitoring sites in the WRAP region with at least three years of valid data and identifies the
six sites selected to demonstrate the apportionment tools. 9-49
Particulate sulfate (a) and nitrate (b) source attribution by region using CAMx modeling for six western remote area
monitoring sites: top left to right Olympic NP, WA; Yellowstone NP, WY; Badlands NP, SD; bottom left to right San
Gorgonio W, CA; Grand Canyon NP, AZ; and Salt Creek W, NM. 9-51
Monthly averaged model predicted organic mass concentration apportioned into primary and anthropogenic and
biogenic secondary PM categories for the Olympic National Park (top) and San Gorgonio Wilderness (bottom)
monitoring sites. From the TSS web site, see Table 9-1. 9-52
Monthly averaged model predicted organic mass concentration apportioned into primary and anthropogenic and
biogenic secondary PM categories for the Yellowstone National Park (top) and Grand Canyon (Hopi Point)
(bottom) monitoring sites. From the TSS web site, see Table 9-1. 9-53
Monthly averaged model predicted organic mass concentration apportioned into primary and anthropogenic and
biogenic secondary PM categories for the Badland National Park (top) and Salt Creek Wilderness (bottom)
monitoring sites. From the TSS web site, see Table 9-1. 9-54
Comparison of carbon concentrations between Seattle (Puget Sound site) and Mt. Rainer (left) and between
Phoenix and Tonto (right) showing the background site concentration (gray) and the urban excess concentration
(black) for total, fossil and contemporary carbon during the summer and winter studies. 9-55
Average contemporary fraction of PM2.5 carbon for the summer (top) and winter (bottom), estimated from
IMPROVE monitoring data (6/04 to 2/06) based on EC/TC ratios. 9-56
Results of the weighted emissions potential tool applied to primary organic carbon emissions (top) and EC
emissions (bottom) for the baseline and projected 2018 emissions inventories for Olympic N.P. 9-58
Results of the weighted emissions potential tool applied to primary organic carbon emissions (top) and EC
emissions (bottom) for the baseline and projected 2018 emissions inventories for San Gorgonio W. 9-59
Results of the weighted emissions potential tool applied to primary organic carbon emissions (top) and EC
emissions (bottom) for the baseline and projected 2018 emissions inventories for Yellowstone N.P. 9-61
Results of the weighted emissions potential tool applied to primary organic carbon emissions (top) and EC
emissions (bottom) for the baseline and projected 2018 emissions inventories for Grand Canyon N.P. 9-62
Results of the weighted emissions potential tool applied to primary organic carbon emissions (top) and EC
emissions (bottom) for the baseline and projected 2018 emissions inventories for Badlands N.P. 9-63
December 2008
xxi
DRAFT-DO NOT QUOTE OR CITE
-------
Figure 9-43. Results of the weighted emissions potential tool applied to primary organic carbon emissions (top) and EC
emissions (bottom) for the baseline and projected 2018 emissions inventories for Salt Creek W. 9-64
Figure 9-44. BRAVO Study haze contributions for Big Bend National Park, TX during a four-month period in 1999. Shown are
impacts by various particulate sulfate sources, as well as the total light extinction level (black line) and Rayleigh or
clear air light scattering. 9-66
Figure 9-45. Maps of spatial patterns for average annual particulate nitrate measurements (top), and for ammonia emissions for
April 2002 from the WRAP emissions inventory (bottom). 9-67
Figure 9-46. Maps of spatial patterns of annual NO (left) and NO2 (right) emissions for 2002 from the WRAP emissions
inventory. 9-68
Figure 9-47. Midwest ammonia monitoring network. 9-69
Figure 9-48. Upwind transport probability fields associated with high particulate nitrate concentrations measured at Toronto,
Canada; Boundary Water Canoe Area, MN; Shenandoah National Park, VA; Lye Brook, VT; and Great Smoky
Mountains National Park, TN. 9-70
Figure 9-49. Trajectory probability fields for periods with high particulate sulfate measured at Underhill, VT and Brigantine, NJ
(shown as white stars) associated with oil-burning trace components (left) and with coal-burning trace components
(right). 9-71
Figure 9-50. Scatter plots of particulate sulfate (left) and particulate nitrate and organic mass (right) versus nephelometer
measured particle light scattering for Acadia National Park, ME. 9-72
Figure 9-51. CMAQ air quality modeling projections of visibility responses on the 20% worst haze days at Great Smoke
National Park, NC (top) and Swanquarter Wilderness, NC (bottom) to 30% reductions from a projected 2009
emission inventory of visibility-reducing pollutants by source category and geographic areas. 9-73
Figure 9-52. The relationship between particle diameter and deposition velocity for particles. Values measured in wind tunnels
by Little and Wiffen (1977) over short grass with wind speed of 2.5 m/s closely approximate the theoretical
distribution determined by Peters and Eiden (1992) for a tall spruce forest. 9-91
Figure 9-53. Relationship of plant nutrients and trace metals with vegetation. Compartments (roman numerals) represent
potential storage sites; whereas arrows (Arabic numerals) represent potential transfer routes. 9-111
Figure 9-54. Atmospheric cycling of aerosols. 9-133
Figure 9-55. Interdependence and feedback between atmospheric aerosol composition, properties, interactions and
transformation, climate and health effects, and sources. 9-134
Figure 9-56. Direct and indirect aerosol effects and major feedback loops in the climate system. 9-135
Figure 9-57. (A) Global mean RFs from the agents and mechanisms discussed in Forster et al. (2007) grouped by agent type. 9-147
Figure 9-58. Components of RF for emissions of principal gases, aerosols and aerosol precursors and other changes. 9-148
Figure 9-59. The transfer of POPs between the major abiotic compartments of the Arctic. Shaded arrows represent
inputs/outputs of POPs to the Arctic. 9-157
December 2008
xxii
DRAFT-DO NOT QUOTE OR CITE
-------
PM ISA Project Team
Executive Direction
Dr. John Vandenberg (Director)—National Center for Environmental Assessment-RTP Division, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Ila Cote (Acting Director)—National Center for Environmental Assessment-RTP Division, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Ms. Debra Walsh (Deputy Director)—National Center for Environmental Assessment-RTP Division, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Mary Ross (Branch Chief)—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Scientific Staff
Dr. Lindsay Wichers Stanek (PM Team Leader)—National Center for Environmental Assessment,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jeffrey Arnold—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Christal Bowman—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. James S. Brown—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Barbara Buckley—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Jean-Jacques Dubois— Oak Ridge Institute for Science and Education, Postdoctoral Research Fellow
to National Center for Environmental Assessment, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Dr. Steven J. Dutton—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Erin Hines—National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Douglas Johns—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Ellen Kirrane—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
December 2008
xxiii
DRAFT-DO NOT QUOTE OR CITE
-------
Dr. Dennis Kotchmar—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Thomas Long—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Thomas Luben—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Qingyu Meng—Oak Ridge Institute for Science and Education, Postdoctoral Research Fellow to
National Center for Environmental Assessment, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Dr. Kristopher Novak—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Joseph Pinto—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Jennifer Richmond-Bryant—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Mary Ross—National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Jason Sacks—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. David Svendsgaard—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. William Wilson—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Lori White—National Institute for Environmental Health Sciences, Research Triangle Park, NC
Technical Support Staff
Mattie Arnold— Senior Environmental Employee Program, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ken Breito— Senior Environmental Employee Program, National Center for Environmental Assessment,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Eleanor Jamison— Senior Environmental Employee Program, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ryan Jones— Oak Ridge Institute for Science and Education, at National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Erica Lee— Oak Ridge Institute for Science and Education, at National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Barbara Liljequist— Senior Environmental Employee Program, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
December 2008
xxiv
DRAFT-DO NOT QUOTE OR CITE
-------
Ellen Lorang— National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Connie Meacham— National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Deborah Wales— National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Erica Wilson— Oak Ridge Institute for Science and Education, at National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Richard Wilson— National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Barbara Wright— Senior Environmental Employee Program, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
December 2008
xxv
DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, Reviewers
Authors
Dr. Lindsay Wichers Stanek (PM Team Leader)—National Center for Environmental Assessment,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jeffrey Arnold—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Christal Bowman—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. James S. Brown—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Barbara Buckley—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Jean-Jacques Dubois—Oak Ridge Institute for Science and Education, Postdoctoral Research Fellow
to National Center for Environmental Assessment, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Dr. Steven J. Dutton—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Erin Hines—National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Douglas Johns—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Ellen Kirrane—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Dennis Kotchmar—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Thomas Long—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Thomas Luben—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Qingyu Meng—Oak Ridge Institute for Science and Education, Postdoctoral Research Fellow to
National Center for Environmental Assessment, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Dr. Kristopher Novak—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Joseph Pinto—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
December 2008
xxvi
DRAFT-DO NOT QUOTE OR CITE
-------
Dr. Jennifer Richmond-Bryant—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Mary Ross—National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Jason Sacks—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Timothy J. Sullivan— E&S Environmental Chemistry, Inc., Corvallis, OR
Dr. David Svendsgaard—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. William Wilson—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Lori White—National Institute for Environmental Health Sciences, Research Triangle Park, NC
Dr. Christy Avery—University of North Carolina, Chapel Hill, NC
Dr. Kathleen Belanger — Center for Perinatal, Pediatric and Environmental Epidemiology, Yale
University, New Haven, CT
Dr. Michelle Bell— School of Forestry & Environmental Studies, Yale University, New Haven, CT
Dr. Matthew J. Campen—Lovelace Respiratory Research Institute, Albuquerque, NM
Dr. Leeland B. Deck— Stratus Consulting, Inc., Washington, DC
Dr. Janneane F. Gent—Center for Perinatal, Pediatric and Environmental Epidemiology, Yale University,
New Haven, CT
Dr. Yuh-Chin Tony Huang—Department of Medicine, Division of Pulmonary Medicine, Duke University
Medical Center, Durham, NC
Dr. Kazuhiko Ito— Nelson Institute of Environmental Medicine, NYU School of Medicine, Tuxedo, NY
Mr. Marc Jackson—Integrated Laboratory Systems, Inc., Research Triangle Park, NC
Dr. Michael Kleinman— Department of Community and Environmental Medicine, University of
California, Irvine
Dr. Marc Pitchford—National Oceanic and Atmospheric Administration, Las Vegas, NV
Dr. Les Recio—Genetic Toxicology Division, Integrated Laboratory Systems, Inc., Research Triangle
Park, NC
Dr. David Quincy Rich—Department of Epidemiology, University of Medicine and Dentistry of New
Jersey, Piscataway, NJ
Dr. Timothy Sullivan— E&S Environmental Chemistry, Inc., Corvallis, OR
Dr. George Thurston—Department of Environmental Medicine, NYU, Tuxedo, NY
Dr. Gregory Wellenius—Cardiovascular Epidemiology Research Unit, Beth Israel Deaconess Medical
Center, Boston, MA
December 2008
xxvii
DRAFT-DO NOT QUOTE OR CITE
-------
Dr. Eric Whitsel—Departments of Epidemiology and Medicine, University of North Carolina, Chapel
Hill, NC
Contributors
Dr. Philip Bromberg—Department of Medicine, University of North Carolina, Chapel Hill, NC
Mr. Michael Burr—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Turhan Carroll—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Rosana Datti—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Neil Frank—Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Jonathan Krug—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Phil Lorang—Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Ms. Christina Miller—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Ms. Irina Mordukhovich—Oak Ridge Institute for Science and Education, at National Center for
Environmental Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Elizabeth Oesterling Owens—Oak Ridge Institute for Science and Education, Postdoctoral Research
Fellow to National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Adam Reff—Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Ms. Vicki Sandiford— Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Mark Schmidt—Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Ms. Angelina Schultz—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Ms. Genee Smith—Oak Ridge Institute for Science and Education, at National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Kurt Susdorf—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
December 2008
xxviii
DRAFT-DO NOT QUOTE OR CITE
-------
Peer Reviewers
Dr. Sara Dubowsky Adar, Department of Epidemiology, University of Washington, Seattle, WA
Dr. Judith Chow, Division of Atmospheric Sciences, Desert Research Institute, Reno, NV
Dr. Dan Costa, Office of Research and Development, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Dr. Robert Devlin, National Health and Environmental Effects Research Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Neil Donahue, Department of Chemical Engineering, Carnegie Mellon University, Pittsburgh, PA
Mr. Tyler Fox, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Mark Frampton, Department of Environmental Medicine, University of Rochester Medical Center,
Rochester, NY
Dr. Jim Gauderman, Department of Environmental Medicine, Department of Preventive Medicine,
University of Southern California, Los Angeles, CA
Dr. Terry Gordon, School of Medicine, New York University, Tuxedo, NY
Mr. Tim Hanley, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Jack Harkema, Department of Pathobiology and Diagnostic Investigation, Michigan State University,
East Lansing, MI
Ms. Beth Hassett-Sipple, Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Amy Herring, Department of Biostatistics, University of North Carolina, Chapel Hill, NC
Dr. Israel Jirak, Department of Meteorology, Embry-Riddle Aeronautical University, Prescott, AZ
Dr. Mike Kleeman, Department of Civil and Environmental Engineering, University of California, Davis,
CA
Dr. Petros Koutrakis, Exposure, Epidemiology and Risk Program, Harvard School of Public Health,
Boston, MA
Dr. Sagar Krupa, Department of Plant Pathology, University of Minnesota, St. Paul, MN
Mr. John Langstafif, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Phil Lorang, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Ms. Karen Martin, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Tom Pace, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
December 2008
xxix
DRAFT-DO NOT QUOTE OR CITE
-------
Dr. Jennifer Peel, Department of Environmental and Radiological Health Sciences, College of Veterinary
Medicine and Biomedical Sciences, Colorado State University, Fort Collins, CO
Mr. Zackary Pekar, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Norm Possiel, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Sanjay Rajagopalan, Division of Cardiovascular Medicine, Ohio State University, Columbus, OH
Mr. Pradeep Rajan, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Venkatesh Rao, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Joanne Rice, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Harvey Richmond, Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Ms. Stefanie Sarnat, Department of Environmental and Occupational Health, Emory University, Atlanta,
GA
Dr. Frances Silverman, Gage Occupational and Environmental Health, University of Toronto, Toronto,
ON
Dr. Barbara Turpin, Department of Environmental Sciences, Rutgers University, New Brunswick, NJ
Mr. Bob Vanderpool, National Exposure Research Laboratory, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Tim Watkins, National Exposure Research Laboratory, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Lewis Weinstock, Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Jason West, Department of Environmental Sciences and Engineering, University of North Carolina,
Chapel Hill, NC
Dr. Antonella Zanobetti, Department of Environmental Health, Harvard School of Public Health, Boston,
MA
December 2008
xxx
DRAFT-DO NOT QUOTE OR CITE
-------
Clean Air Scientific Advisory Committee
for Particulate Matter NAAQS
Chairperson
Dr. Jonathan Samet, Department of Epidemiology, Bloomberg School of Public Health, Johns Hopkins
University, Baltimore, MD
Members
Dr. Lowell Ashbaugh, Crocker Nuclear Lab, University of California, Davis, CA
Dr. Ed Avol, Preventive Medicine, Keck School of Medicine, University of Southern California, Los
Angeles, CA
Dr. Joseph Brain*, Department of Environmental Health, Harvard School of Public Health, Harvard
University, Boston, MA
Dr. Wayne Cascio, Brody School of Medicine at East Carolina University, Greenville, NC
Dr. Ellis B. Cowling*, Colleges of Natural Resources and Agriculture and Life Sciences, North Carolina
State University, Raleigh, NC
Dr. James Crapo*, Department of Medicine, National Jewish Medical and Research Center, Denver, CO
Dr. Douglas Crawford-Brown, Department of Environmental Sciences and Engineering, University of
North Carolina at Chapel Hill, Chapel Hill, NC
Dr. H. Christopher Frey*, Department of Civil, Construction and Environmental Engineering, College of
Engineering, North Carolina State University, Raleigh, NC
Dr. David Grantz, Botany and Plant Sciences and Air Pollution Research Center, Riverside Campus and
Kearney Agricultural Center, University of California, Parlier, CA
Dr. Joseph Helble, Thayer School of Engineering, Dartmouth College, Hanover, NH
Dr. Rogene Henderson**, Lovelace Respiratory Research Institute, Albuquerque, NM
Dr. Philip Hopke, Department of Chemical Engineering, Clarkson University, Potsdam, NY
Dr. Donna Kenski*, Lake Michigan Air Directors Consortium, Rosemont, IL
Dr. Morton Lippmann, Nelson Institute of Environmental Medicine, New York University School of
Medicine, Tuxedo, NY
Dr. Helen Suh Macintosh, Environmental Health, School of Public Health, Harvard University, Boston,
MA
Dr. William Malm, National Park Service Air Resources Division, Cooperative Institute for Research in
the Atmosphere, Colorado State University, Fort Collins, CO
December 2008
xxxi
DRAFT-DO NOT QUOTE OR CITE
-------
Mr. Charles Thomas (Tom) Moore, Jr., Cooperative Institute for Research in the Atmosphere, Colorado
State University, Fort Collins, CO
Dr. Robert F. Phalen, Center for Occupation & Environment Health, College of Medicine, Department of
Community and Environmental Medicine, Air Pollution Health Effects Laboratory, University of
California Irvine, Irvine, CA
Dr. Kent Pinkerton, Center for Health and the Environment, University of California, Davis, CA
Mr. Richard L. Poirot, Air Pollution Control Division, Department of Environmental Conservation,
Vermont Agency of Natural Resources, Waterbury, VT
Dr. Armistead (Ted) Russell*, Department of Civil and Environmental Engineering , Georgia Institute of
Technology, Atlanta, GA
Dr. Frank Speizer, Channing Laboratory, Harvard Medical School, Boston, MA
Dr. Sverre Vedal, Department of Environmental and Occupational Health Sciences, School of Public
Health and Community Medicine, University of Washington, Seattle, WA
* Members of the statutory Clean Air Scientific Advisory Committee (CAS AC) appointed by the EPA
Administrator.
**As immediate past CASAC Chair, Dr. Henderson is invited to participate in CASAC advisory activities
for FY 2009.
December 2008
xxxii
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Chapter 1. Introduction
The first external review draft Integrated Science Assessment (ISA) is a concise review, synthesis,
and evaluation of the most policy-relevant science, and communicates critical science judgments relevant
to the National Ambient Air Quality Standards (NAAQS) review. As such, the ISA forms the scientific
foundation for the review of the primary (health-based) and secondary (welfare-based) NAAQS for
particulate matter (PM). The ISA accurately reflects "the latest scientific knowledge useful in indicating
the kind and extent of identifiable effects on public health which may be expected from the presence of
[a] pollutant in ambient air" (42 U.S.C. 7408). Key information and judgments formerly contained in the
Air Quality Criteria Document (AQCD) for PM are incorporated in this assessment. Additional details of
the pertinent scientific literature published since the last review, as well as selected older studies of
particular interest, are included in a series of annexes. This ISA thus serves to update and revise the
information available at the time of the previous review of the NAAQS for PM in 2006.
The Integrated Review Plan for the National Ambient Air Quality Standards for Particulate Matter
identifies a series of policy-relevant questions that provide a framework for this review of the scientific
evidence (U.S. EPA, 2008b). These questions frame the entire review of the NAAQS for PM, and thus are
informed by both science and policy considerations. The ISA organizes and presents the scientific
evidence such that, when considered along with findings from risk analyses and policy considerations,
will help the EPA address these questions during the NAAQS review for PM. Briefly, the focus of this
assessment will be on scientific evidence that is most relevant to the following:
¦ What new scientific evidence is available to better understand the relationship between health
effects and short- or long-term exposure to PM? To what extent has scientific evidence
improved our understanding of the nature and magnitude of visibility, climate, and ecosystem
responses to PM?
¦ What evidence is available from recent studies focused on specific size fractions, chemical
components, sources, or environments (e.g., urban and non-urban areas) of PM to inform our
understanding of the nature of PM exposures that are linked to various health or public
welfare effects?
December 2008
1-1
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
¦ At what levels of PM exposure are health or welfare effects observed? What is the nature of
the dose-response relationships of PM for the various effects evaluated?
¦ To what extent is key evidence becoming available that could inform our understanding of
subpopulations that are particularly susceptible or vulnerable to PM exposures1?
¦ To what extent have important uncertainties identified in the last review been reduced? Have
new uncertainties emerged?
1.1. Legislative Requirements
Two sections of the Clean Air Act (CAA) govern the establishment and revision of the NAAQS.
Section 108 (42 U.S.C. 7408) directs the Administrator to identify and list "air pollutants" that "in his
judgment, may reasonably be anticipated to endanger public health and welfare" and whose "presence...
in the ambient air results from numerous or diverse mobile or stationary sources" and to issue air quality
criteria for those that are listed (42 U.S.C. 7408). Air quality criteria are intended to "accurately reflect the
latest scientific knowledge useful in indicating the kind and extent of identifiable effects on public health
or welfare which may be expected from the presence of [a] pollutant in ambient air..." 42 U.S.C. 7408(b).
Section 109 (42 U.S.C. 7409) directs the Administrator to propose and promulgate "primary" and
"secondary" NAAQS for pollutants listed under Section 108. 42 U.S.C. 7409(a). Section 109(b)(1)
defined a primary standard as one "the attainment and maintenance of which in the judgment of the
Administrator, based on such criteria and allowing an adequate margin of safety, are requisite to protect
the public health."2 42 U.S.C. 7409(b)(1). A secondary standard, as defined in Section 109(b)(2), must
"specify a level of air quality the attainment and maintenance of which, in the judgment of the
Administrator, based on such criteria, is required to protect the public welfare from any known or
1 "Susceptibility" refers to innate (e.g., genetic or developmental) or acquired (e.g., age, disease, or smoking) factors that make individuals more
likely to experience effects with exposure to PM. "Vulnerability" refers to PM-related effects due to factors including socioeconomic status
(e.g., reduced access to health care) or particularly elevated exposure levels.
2 The legislative history of section 109 indicates that a primary standard is to be set at "the maximum permissible ambient air level.. .which will
protect the health of any [sensitive] group of the population," and that for this purpose "reference should be made to a representative sample of
persons comprising the sensitive group rather than to a single person in such a group" [S. Rep. No. 91-1196, 91st Cong., 2d Sess. 10 (1970)].
December 2008
1-2
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
anticipated adverse effects associated with the presence of [the] pollutant in the ambient air."1
42 U.S.C. 7409(b)(2).
The requirement that primary standards include an adequate margin of safety was intended to
address uncertainties associated with inconclusive scientific and technical information available at the
time of standard setting. It was also intended to provide a reasonable degree of protection against hazards
that research has not yet identified. See Lead Industries Association v. EPA, 647 F.2d 1130, 1154
(D.C. Cir 1980), cert, denied, 449 U.S. 1042 (1980); American Petroleum Institute v. Costle, 665 F.2d
1176, 1186 (D.C. Cir. 1981), cert, denied, 455 U.S. 1034 (1982). Both kinds of uncertainties are
components of the risk associated with pollution at levels below those at which human health effects can
be said to occur with reasonable scientific certainty. Thus, in selecting primary standards that include an
adequate margin of safety, the Administrator is seeking not only to prevent pollution levels that have been
demonstrated to be harmful but also to prevent lower pollutant levels that may pose an unacceptable risk
of harm, even if the risk is not precisely identified as to nature or degree.
In selecting a margin of safety, the EPA considers such factors as the nature and severity of the
health effects involved, the size of the sensitive population(s) at risk, and the kind and degree of the
uncertainties that must be addressed. The selection of any particular approach to providing an adequate
margin of safety is a policy choice left specifically to the Administrator's judgment. See Lead Industries
Association v. EPA, supra, 647 F.2d 1161-62.
In setting standards that are "requisite" to protect public health and welfare, as provided in
Section 109(b), the Administrator's task is to establish standards that are neither more nor less stringent
than necessary. In so doing, EPA may not consider the costs of implementing the standards. See generally
Whitman v. American Trucking Associations, 531 U.S. 457, 465-472, 475-76 (2001).
Section 109(d)(1) requires that "not later than December 31, 1980, and at 5-year
intervals thereafter, the Administrator shall complete a thorough review of the criteria published under
Section 108 and the national ambient air quality standards... and shall make such revisions in such criteria
and standards and promulgate such new standards as may be appropriate..." 42 U.S.C. 7409(d)(1).
Section 109(d)(2) requires that an independent scientific review... committee "shall complete a review of
the criteria and the national primary and secondary ambient air quality standards... and shall recommend
to the Administrator any new standards and revisions of existing criteria and standards as may be
1 Welfare effects as defined in Section 302(h) [42 U.S.C. 7602(h)] include, but are not limited to, "effects on soils, water, crops, vegetation,
man-made materials, animals, wildlife, weather, visibility and climate, damage to and deterioration of property, and hazards to transportation,
December 2008
1-3
DRAFT-DO NOT QUOTE OR CITE
-------
1 appropriate..." 42 U.S.C. 7409(d)(2). Since the early 1980s, this independent review function has been
2 performed by the Clean Air Scientific Advisory Committee (CASAC).
1.2. History of Reviews of the NAAQS for PM
3 PM is the generic term for a broad class of chemically and physically diverse substances that exist
4 as discrete particles (liquid droplets or solids) over a wide range of sizes. Particles originate from a variety
5 of anthropogenic stationary and mobile sources as well as from natural sources. Particles may be emitted
6 directly or formed in the atmosphere by transformations of gaseous emissions such as sulfur oxides
7 (SOx), nitrogen oxides (NOx), and volatile organic compounds (VOC). The chemical and physical
8 properties of PM vary greatly with time, region, meteorology, and source category, thus complicating the
9 assessment of health and welfare effects. Table 1-1 summarizes the NAAQS that have been promulgated
10 for PM to date. These reviews are briefly described below, and further details are provided in the
11 Integrated Review Plan (U.S. EPA, 2008b).
as well as effects on economic values and on personal comfort and well-being."
December 2008
1-4
DRAFT-DO NOT QUOTE OR CITE
-------
Table 1-1. Summary of NAAQS promulgated for PM, 1971-2006.
Final Rule
Indicator
Avg
Time
Level
Form
1971 (36 FR 8186)
TSP (Total
Suspended
Particulates)
24-h
260 (jg/rn3 (primary)
150 (jg/rn3 (secondary)
Not to be exceeded more than once per year
Annual
75 (jg/rn3 (primary)
Annual average
1987 (52 FR 24634)
PM10
24-h
150 (jg/m3
Not to be exceeded more than once per year on average over a 3-year
period
Annual
50 (jg/m3
Annual arithmetic mean, averaged over 3 years
1997 (62 FR 38652)
PM2.5
24-h
65 (jg/m3
98th percentile, averaged over 3 years
Annual
15 (jg/m3
Annual arithmetic mean, averaged over 3 years1
PM10
24-h
150 (jg/m3
Initially promulgated 99th percentile, averaged over 3 years; when 1997
standards were vacated in 1999, the form of 1987 standards remained in
place (not to be exceeded more than once per year on average over a
3-year period)
Annual
50 (jg/m3
Annual arithmetic mean, averaged over 3 years
2006(71 FR 61144)
PM2.5
24-h
35 (jg/m3
98th percentile, averaged over 3 years
Annual
15 (jg/m3
Annual arithmetic mean, averaged over 3 years2
PM10
24-h
150 (jg/m3
Not to be exceeded more than once per year on average over a 3-year
period
Note: When not specified, primary and secondary standards are identical.
1 EPA first established NAAQS for PM in 1971 (36 FR 8186, April 30, 1971), based on the original
2 criteria document (NAPA, 1969). The reference method specified for determining attainment of the
3 original standards was the high-volume sampler, which collects PM up to a nominal size of 25 to 45
4 micrometers ((.im) (referred to as total suspended particulates or TSP). The primary standards (measured
5 by the indicator TSP) were 260 (ig/m3, 24-h average, not to be exceeded more than once per year, and
6 75 (ig/m3, annual geometric mean. The secondary standard was 150 (ig/m3, 24-h average, not to be
7 exceeded more than once per year. In October 1979 (44 FR 56730, October 2, 1979), EPA announced the
8 first periodic review of the air quality criteria and NAAQS for PM, and significant revisions to the
9 original standards were promulgated in 1987 (52 FR 24634, July 1, 1987). In that decision, EPA changed
1 The level of the 1997 annual PM2.5 standard was to be compared to measurements made at the community-oriented monitoring site recording
the highest level, or, if specific constraints were met, measurements from multiple community-oriented monitoring sites could be averaged
("spatial averaging"). These criteria and constraints were intended to ensure that spatial averaging would not result in inequities in the level of
protection afforded by the PM2.5 standards. Community-oriented monitoring sites were specified to be consistent with the intent that a spatially
averaged annual standard provide protection for persons living in smaller communities, as well as those in larger population centers.
2 In the revisions to the PM NAAQS finalized in 2006, EPA tighten the constraints on the spatial averaging criteria by further limiting the
conditions under which some areas may average measurements from multiple community-oriented monitors to determine compliance (see
71 FR 61165-61167, October 17, 2006).
December 2008
1-5
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
the indicator for particles from TSP to PM10, the latter including particles with a mean aerodynamic
diameter1 less than or equal to 10 (.im, which delineated that subset of inhalable particles small enough to
penetrate to the thoracic region (including the tracheobronchial and alveolar regions) of the respiratory
tract (referred to as thoracic particles). EPA also revised the level and form of the primary standards by
(1) replacing the 24-h TSP standard with a 24-h PMi0 standard of 150 (ig/m3 with no more than one
expected exceedence per year; and (2) replacing the annual TSP standard with a PMi0 standard of
50 (.ig/ni3. annual arithmetic mean, averaged over three years.
The secondary standard was revised by replacing it with 24-h and annual standards identical in all
respects to the primary standards. The revisions also included a new reference method for the
measurement of PMi0 in the ambient air and rules for determining attainment of the new standards. On
judicial review, the revised standards were upheld in all respects. Natural Resources Defense Council v.
Administrator, 902 F. 2d 962 (D.C. Cir. 1990, cert, denied, 498 U.S. 1082, 1991).
In April 1994, EPA announced its plans for the second periodic review of the air quality criteria and
NAAQS for PM, and promulgated significant revisions to the NAAQS in 1997 (62 FR 38652, July 18,
1997). In that decision, EPA revised the PM NAAQS in several respects. While EPA determined that the
PM NAAQS should continue to focus on PMi0, EPA also determined that the fine and coarse1 fractions of
PM10 should be considered separately. The Administrator's decision to modify the standards was based on
evidence that serious health effects were associated with short- and long-term exposure to fine particles in
areas that met the existing PM10 standards. The EPA added new standards, using PM2 5 as the indicator for
fine particles (with PM2 5 referring to particles with a nominal mean aerodynamic diameter less than or
equal to 2.5 and using PM10 as the indicator for purposes of regulating the coarse fraction of PM10
(referred to as thoracic coarse particles or coarse-fraction particles; generally including particles with a
nominal mean aerodynamic diameter greater than 2.5 |_im and less than or equal to 10 |_im. or PMi0_2.5).
The EPA established two new PM2 5 standards: an annual standard of 15 (ig/m3, based on the 3-year
average of annual arithmetic mean PM2 5 concentrations from single or multiple community-oriented
monitors; and a 24-h standard of 65 (ig/m3, based on the 3-year average of the 98th percentile of 24-h
PM2 5 concentrations at each population-oriented monitor within an area. Also, EPA established a new
reference method for the measurement of PM2 5 in the ambient air and adopted rules for determining
1 The more precise term is 50 percent cut point or 50 percent diameter (dso). This is the aerodynamic particle diameter for which the efficiency of
particle collection is 50 percent. Larger particles are not excluded altogether, but are collected with substantially decreasing efficiency and
smaller particles are collected with increasing (up to 100 percent) efficiency.
December 2008
1-6
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
attainment of the new standards. To continue to address thoracic coarse particles, EPA retained the annual
PM10 standard, while revising the form, but not the level, of the 24-h PM10 standard to be based on the
99th percentile of 24-h PM10 concentrations at each monitor in an area. The EPA revised the secondary
standards by making them identical in all respects to the primary standards.
Following promulgation of the 1997 PM NAAQS, petitions for review were filed by a large
number of parties, addressing a broad range of issues. In May 1999, a three-judge panel of the U.S. Court
of Appeals for the District of Columbia Circuit issued an initial decision that upheld EPA's decision to
establish fine particle standards, holding that "the growing empirical evidence demonstrating a
relationship between fine particle pollution and adverse health effects amply justifies establishment of
new fine particle standards." American Trucking Associations v. EPA (175 F. 3d 1027, 1055-56 [D.C. Cir.
1999; rehearing granted in part and denied in part, 195 F. 3d 4 [D.C. Cir. 1999]), affirmed in part and
reversed in part, Whitman v. American Trucking Associations (531 U.S. 457, 2001). The Panel also found
"ample support" for EPA's decision to regulate coarse particle pollution, but vacated the 1997 PMi0
standards, concluding that EPA had not provided a reasonable explanation justifying use of PMi0 as an
indicator for coarse particles (175 F. 3d at 1054-55). Pursuant to the court's decision, EPA removed the
vacated 1997 PMi0 standards. The pre-existing 1987 PMi0 standards remained in place (65 FR 80776,
December 22, 2000). The Court also upheld EPA's determination not to establish more stringent
secondary standards for fine particles to address effects on visibility (175 F. 3d 1027).
More generally, the panel held (over one judge's dissent) that EPA's approach to establishing the
level of the standards in 1997, both for PM and for ozone (03) NAAQS promulgated on the same day,
effected "an unconstitutional delegation of legislative authority" (Id. at 1034-40). Although the panel
stated that "the factors EPA uses in determining the degree of public health concern associated with
different levels of ozone and PM are reasonable," it remanded the rule to EPA, stating that when EPA
considers these factors for potential non-threshold pollutants "what EPA lacks is any determinate criterion
for drawing lines" to determine where the standards should be set. Consistent with EPA's long-standing
interpretation and D.C. Circuit precedent, the panel also reaffirmed prior rulings holding that in setting
NAAQS EPA is "not permitted to consider the cost of implementing those standards" (Id. at 1040-41).
Both sides filed cross appeals on these issues to the U.S. Supreme Court, and the Court granted
certiorari. In February 2001, the Supreme Court issued a unanimous decision upholding EPA's position
on both the constitutional and cost issues. Whitman v. American Trucking Associations, 531 U.S. 457,
1 See definitions of "fine" and "coarse" particles in Section 3.2.
December 2008
1-7
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
464, 475-76. On the constitutional issue, the Court held that the statutory requirement that NAAQS be
"requisite" to protect public health with an adequate margin of safety sufficiently guided EPA's discretion,
affirming EPA's approach of setting standards that are neither more nor less stringent than necessary. The
Supreme Court remanded the case to the Court of Appeals for resolution of any remaining issues that had
not been addressed in that court's earlier rulings. Id. at 475-76. In March 2002, the Court of Appeals
rejected all remaining challenges to the standards, holding under the traditional standard of judicial
review that PM2.5 standards were reasonably supported by the administrative record and were not
"arbitrary and capricious" American Trucking Associations v. EPA, 283 F. 3d 355, 369-72 (D.C. Cir.
2002).
In October 1997, EPA published its plans for the third periodic review of the air quality criteria and
NAAQS for PM (62 FR 55201). After CASAC and public review, EPA's NCEA finalized the 2004 PM
AQCD (U.S. EPA, 2004). The final Office of Air Quality Planning and Standards (OAQPS) Staff Paper
(U.S. EPA, 2005b), took into account the advice and recommendations of CASAC and public comments
received on the earlier drafts of this document and presented additional advice and recommendations
submitted by CASAC to the Administrator.
On December 20, 2005, EPA announced its proposed decision to revise the NAAQS for PM (71 FR
2620; hereafter "proposal"). In the proposal, EPA identified proposed revisions, based on the air quality
criteria for PM, and solicited public comments on alternative primary and secondary standards. EPA
proposed to revise the level of the 24-h PM25 standard to 35 (.ig/nr1 to provide increased protection against
health effects associated with short-term PM2 5 exposures, including premature mortality and increased
hospital admission and emergency room visits and to retain the level of the annual PM2 5 standard at
15 (ig/m3, continuing protection against health effects associated with long-term exposure including
premature mortality and development of chronic respiratory disease. With regard to the primary standards
for PMio, EPA proposed to revise the 24-h PMi0 standard in part by establishing a new indicator for
thoracic coarse particles (particles generally between 2.5 and 10 |_im in PMi0.2 5 diameter), qualified so as
to include any ambient mix of PMi0.2.5 that was dominated by resuspended dust from high density traffic
on paved roads and PM generated by industrial sources and construction sources, and proposed to exclude
any ambient mix of PMi0.2.5 that was dominated by rural windblown dust and soils and PM generated by
agricultural and mining sources. The EPA proposed to set a 24-h PMi0.2 5 standard at a level of 70 (ig/m3
to continue to provide a level of protection against health effects associated with short-term exposure
(including hospital admissions for cardiopulmonary diseases, increased respiratory symptoms and
possibly premature mortality) generally equivalent to the level of protection provided by the existing 24-h
PM10 standard. Also, EPA proposed to revoke, upon finalization of a primary 24-h standard for PMi0.2 5,
the 24-h PMi0 standard as well as the annual PMi0 standard. EPA proposed to revise the secondary
December 2008
1-8
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
standards by making them identical to the suite of proposed primary standards for fine and coarse
particles, providing protection against PM-related public welfare effects including visibility impairment,
effects on vegetation and ecosystems, and materials damage and soiling. EPA also solicited comment on
adding a new sub-daily PM2 5 secondary standard to address visibility impairment in urban areas. CAS AC
provided additional advice to EPA in a letter to the Administrator requesting reconsideration of CASAC's
recommendations for both the primary and secondary PM2 5 standards as well as standards for thoracic
coarse particles (Henderson, 2006).
On September 21, 2006, EPA announced its final decisions to revise the primary and secondary
NAAQS for PM to provide increased protection of public health and welfare, respectively (71 FR 61144).
With regard to the primary and secondary standards for fine particles, EPA revised the level of the 24-h
PM2 5 standard to 35 (ig/m3, retained the level of the annual PM2 5 annual standard at 15 (ig/m3, and
revised the form of the annual PM2 5 standard by narrowing the constraints on the optional use of spatial
averaging. With regard to the primary and secondary standards for PMi0, EPA retained the 24-h PMi0
standard at 150 (ig/m3 and revoked the annual standard because available evidence generally did not
suggest a link between long-term exposure to current ambient levels of coarse particles and health or
welfare effects.
1.3. Document Development
EPA initiated the current formal review of the NAAQS for PM on June 28, 2007 with a call for
information from the public (72 FR 35462). In addition to the call for information, publications were
identified through an ongoing literature search process that includes extensive computer database mining
on specific topics. Additional publications were identified by EPA scientists in a variety of disciplines by
combing through relevant, peer-reviewed scientific literature obtained through these ongoing literature
searches, reviewing previous EPA reports, and a review of reference lists from key publications; studies
were also identified in the course of CASAC and public review.
All relevant epidemiologic, human clinical, and animal toxicological studies, including those
related to exposure-response relationships, mode(s) of action (MOA), or susceptible or vulnerable
subpopulations, and ecological or welfare effects studies published since the last review were considered.
Added to the body of research were EPA's analyses of air quality and emissions data, studies on
atmospheric chemistry, transport, and fate of these emissions, as well as issues related to exposure to PM.
Further information was acquired from consultation with content and area experts and the public.
In the 2008 NOx-SOx ISA (U.S. EPA, 2008e), EPA focused on ecological effects related to the
deposition of nitrogen (N)- and sulfur (S)-containing compounds, including particle-phase compounds
December 2008
1-9
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
(e.g., nitrates and sulfates) primarily including effects from acidification and N-nutrient enrichment and
eutrophication. In this draft ISA, EPA focused on recent data on direct welfare effects of particle-phase
NOx and SOx in the ambient air — primarily visibility impairment, damage to materials, and positive and
negative climate interactions.
1.4. Document Organization
The ISA is composed of nine chapters. This introductory chapter presents background information,
and provides an overview of EPA's framework for making causal judgments. Key findings and
conclusions from the atmospheric sciences, ambient air data analyses, exposure assessment, dosimetry,
and health effects for consideration in the review of the NAAQS for PM, including judgments on
causality for the health effects of PM exposure, are presented in Chapter 2. More detailed summaries,
evaluations and integration of the evidence are included in Chapters 3 through 8.
Chapter 3 highlights key concepts or issues relevant to understanding the atmospheric chemistry,
sources, and exposure of PM following a "source-to-exposure" paradigm. Chapter 4 summarizes key
concepts and recent findings on the dosimetry of PM and Chapter 5 discusses possible pathways and
MOAforthe effects of PM. Chapters 6 and 7 evaluate and integrate epidemiologic, human clinical, and
animal toxicological information relevant to the review of the primary NAAQS for PM. Effects related to
short-term exposures to PM are the focus of Chapter 6. Chapter 7 evaluates evidence related to long-term
exposures to PM. Chapters 6 and 7 are organized by health outcome categories, such as cardiovascular or
respiratory effects, and each section includes effects of the various types of PM studied. For each health
outcome category, summary sections then integrate the findings to draw conclusions on the evidence for
the main size classes of PM (i.e., PMi0, PM2.5, PM10-2.5, and ultrafine particles). Chapter 6 also includes a
summary and synthesis of the recent evidence on various health effects related to short-term exposure to
different components or sources of PM. Chapter 8 evaluates evidence related to the public health impact
of ambient PM exposure, including potentially susceptible and vulnerable population groups.
Chapter 9 evaluates ecological and welfare effects evidence that is relevant to the review of the
secondary NAAQS for PM. That chapter includes consideration of evidence on visibility impairment,
materials damage, effects of PM on climate, and ecological effects of PM that were not addressed in the
2008 NOx-SOx ISA (U.S. EPA, 2008e); the chapter also includes the integrative synthesis of the evidence
and presents key conclusions and scientific judgments regarding causality for welfare and ecological
effects of PM.
December 2008
1-10
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
A series of annexes supplement this ISA. The annexes provide additional details of the pertinent
literature published since the last review, as well as selected older studies of particular interest. These
annexes contain information on:
¦ atmospheric chemistry of PM as well as the sampling and analytic methods for measurement
of PM (Annex A)
¦ concentrations, emissions, sources and human exposure to PM (Annex A)
¦ studies on the dosimetry of PM (Annex B)
¦ human clinical studies of health effects related to exposure to PM (Annex D)
¦ toxicological studies of health effects in laboratory animals (Annex C); and
¦ epidemiologic studies of health effects from short- and long-term exposure to PM (Annex E)
Within the Annexes, detailed information about methods and results of health studies is
summarized in tabular format, and generally includes information about: concentrations of PM and
averaging times; study methods employed; results and comments; and quantitative results for
relationships between effects and exposure to PM.
1.5. EPA Framework for Causal Determination
The EPA has developed a consistent and transparent basis to evaluate the causal nature of air
pollution-induced health or environmental effects. The framework described below establishes uniform
language concerning causality and brings more specificity to the findings. It drew standardized language
from across the federal government and wider scientific community, especially from the recent National
Academy of Sciences (NAS) Institute of Medicine (IOM) document, Improving the Presumptive
Disability Decision-Making Process for Veterans (IOM, 2008), the most recent comprehensive work on
evaluating causality.
This introductory section focuses on the evaluation of health effects evidence, particularly
epidemiologic study results, since this is a crucial element of the PM ISA. While focusing on human
health outcomes, the concepts are also generally relevant to causality determination for welfare effects.
This section:
¦ describes the kinds of scientific evidence used in establishing a general causal relationship
between exposure and health effects;
¦ defines cause, in contrast to statistical association;
December 2008
1-11
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
¦ discusses the sources of evidence necessary to reach a conclusion about the existence of a
causal relationship;
¦ highlights the issue of multifactorial causation;
¦ identifies issues and approaches related to uncertainty; and
¦ provides a framework for classifying and characterizing the weight of evidence in support of
a general causal relationship.
Approaches to assessing the separate and combined lines of evidence (e.g., epidemiologic, human
clinical, and animal toxicological studies) have been formulated by a number of regulatory and science
agencies, including the IOM of the NAS (IOM, 2008), International Agency for Research on Cancer
(IARC, 2006), EPA Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), Centers for Disease
Control and Prevention (CDC, 2004), and National Acid Precipitation Assessment Program (NAPAP)
(1991). These formalized approaches offer guidance for assessing causality. The frameworks are similar
in nature, although adapted to different purposes, and have proven effective in providing a uniform
structure and language for causal determinations. Moreover, these frameworks have supported
decision-making under conditions of uncertainty.
1.5.1. Scientific Evidence Used in Establishing Causality
Causality determinations are based on the evaluation and synthesis of evidence from across
scientific disciplines; the type of evidence that is most important for such determinations will vary by
pollutant or assessment. The most compelling evidence of a causal relationship between pollutant
exposures and human health effects comes from human clinical studies. This type of study experimentally
evaluates the health effects of administered exposures in human volunteers under highly-controlled
laboratory conditions.
In epidemiologic or observational studies of humans, the investigator does not control exposures or
intervene with the study population. Broadly, observational studies can describe associations between
exposures and effects. These studies fall into several categories: cross-sectional, prospective cohort, and
time-series studies. "Natural experiments" offer the opportunity to investigate changes in health with a
change in exposure; these include comparisons of health effects before and after a change in population
exposures, such as closure of a pollution source.
Experimental animal data complement the clinical and observational data; these studies can help
characterize effects of concern, exposure-response relationships, susceptible subpopulations and MOAs.
December 2008
1-12
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
In the absence of clinical or epidemiologic data, animal data alone may be sufficient to support a likely
causal determination, assuming that humans respond similarly to the experimental species.
1.5.2. Association and Causation
"Cause" is a significant, effectual relationship between an agent and an effect on health or public
welfare. "Association" is the statistical dependence among events, characteristics, or other variables. An
association is prima facie evidence for causation; alone, however, it is insufficient proof of a causal
relationship between exposure and disease. Unlike an association, a causal claim supports the creation of
counterfactual claims; that is, a claim about what the world would have been like under different or
changed circumstances (IOM, 2008). Much of the newly available health information evaluated in this
ISA comes from epidemiologic studies that report a statistical association between ambient exposure and
health outcome.
Many of the health and environmental outcomes reported in these studies have complex etiologies.
Diseases such as asthma, coronary heart disease or cancer are typically initiated by a web of multiple
agents. Outcomes depend on a variety of factors, such as age, genetic susceptibility, nutritional status,
immune competence, and social factors (Gee and Payne-Sturges, 2004; IOM, 2008). Effects on
ecosystems are often also multifactorial with a complex web of causation. Further, exposure to a
combination of agents could cause synergistic or antagonistic effects. Thus, the observed risk represents
the net effect of many actions and counteractions.
1.5.3. Evaluation of Evidence for Going beyond Association to
Causation
Moving from association to causation involves elimination of alternative explanations for the
association. In estimating the causal influence of an exposure on health or environmental effects, it is
recognized that scientific findings incorporate uncertainty. Uncertainty can be defined as a state of having
limited knowledge where it is impossible to exactly describe an existing state or future outcome; the lack
of knowledge about the correct value for a specific measure or estimate. Uncertainty characterization and
uncertainty assessment are two activities that lead to different degrees of sophistication in describing
uncertainty. Uncertainty characterization generally involves a qualitative discussion of the thought
processes that lead to the selection and rejection of specific data, estimates, scenarios, etc. The uncertainty
assessment is more quantitative. The process begins with simpler measures (e.g., ranges) and simpler
analytical techniques and progresses, to the extent needed to support the decision for which the
December 2008
1-13
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
assessment is conducted, to more complex measures and techniques. Data will not be available for all
aspects of an assessment and those data that are available may be of questionable or unknown quality. In
these situations, evaluation of uncertainty can include professional judgment or inferences based on
analogy with similar situations. The net result is that the assessments will be based on a number of
assumptions with varying degrees of uncertainty.
Uncertainties commonly encountered in evaluating health evidence for the criteria air pollutants are
outlined below for epidemiologic and experimental studies. Various approaches to characterizing
uncertainty include classical statistical methods, sensitivity analysis, or probabilistic uncertainty analysis,
in order of increasing complexity and data requirements. The ISA generally evaluates uncertainties
qualitatively in assessing the evidence from across studies; in some situations quantitative analysis
approaches, such as meta-regression may be used.
Controlled human exposure studies evaluate the effects of exposures to a variety of pollutants in a
highly controlled laboratory setting. Also referred to as human clinical studies, these experiments allow
investigators to expose subjects to fixed concentrations of air pollutants under carefully regulated
environmental conditions and activity levels. Controlled human exposures to PM typically involve
exposing subjects either at rest or while engaged in intermittent exercise in a whole-body exposure
chamber, although mouthpiece and facemask systems can also be used. A variety of different types of
particles are used in these studies including ambient outdoor particles, concentrated ambient particles
(CAPs), diesel exhaust (DE) from a diesel engine, wood smoke (WS) generated in a wood stove,
laboratory generated surrogate particles (e.g., elemental carbon [EC]or zinc oxide [ZnO]), or particles
collected on a filter, resuspended in saline, and administered either through instillation or inhalation
(aerosolized and delivered using a nebulizer). The recovery of particles on filters is variable and some
components, such as organics, may be too volatile to be collected. Exposures to artificially generated
particles may provide important information on the health effects of PM, but are not truly representative
of ambient air pollution particles. The direct exposure of humans to ambient air pollution particles may be
complicated by factors that cannot be controlled such as co-exposures to other air pollutants (e.g., 03,
S02, and N02). In concentrating ambient particulates, gaseous copollutants are not proportionately
concentrated and interactions between PM and the copollutants cannot be investigated unless the latter are
re-introduced. These limitations as well as daily variability in concentration and composition can make it
difficult to compare the results from human clinical studies employing particles from different sources.
In some instances, controlled human exposure studies can also be used to characterize
concentration-response relationships at pollutant concentrations relevant to ambient conditions.
Controlled human exposures are typically conducted using a randomized crossover design with subjects
exposed both to PM and a clean air control. In this way, subjects serve as their own controls, effectively
December 2008
1-14
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
controlling for many potential confounders. However, human clinical studies are limited by a number of
factors including a small sample size and short exposure time. The repetitive nature of ambient PM
exposures may lead to cumulative health effects, but this type of exposure is not practical to replicate in a
laboratory setting. In addition, although subjects do serve as their own controls, personal exposure to
pollutants in the hours and days preceding the controlled exposures may vary significantly between and
within individuals. Finally, human clinical studies require investigators to adhere to stringent health
criteria for a subject to be included in the study, and therefore the results cannot necessarily be
generalized to an entire population. Although some human clinical studies have included health
comprised individuals such as asthmatics or individuals with chronic obstructive pulmonary disease
(COPD) or coronary artery disease, these individuals must also be relatively healthy and do not represent
the most sensitive individuals in the population. Thus, a lack of observation of effects from human
clinical studies does not necessarily mean that a causal relationship does not exist. While human clinical
studies provide important information on the biological plausibility of associations observed between air
pollutant exposure and health outcomes in epidemiologic studies, observed effects in these studies may
underestimate the response in certain subpopulations.
Epidemiologic studies provide important information on the associations between health effects
and exposure of human populations to ambient air pollution. These studies also help to identify
susceptible or vulnerable subgroups and associated risk factors. There are important methodological
issues that to be considered in evaluating results from air pollution epidemiologic studies, especially the
potential for confounding and/or effect modification; and exposure measurement error.
Scientific judgment is needed regarding sources and magnitude of potential confounding by
covariates, together with judgment about how well the existing constellation of study designs, results, and
analyses address this potential threat to inferential validity. One key consideration is evaluation of the
potential contribution of PM to health effects, when it is a component of a complex air pollutant mixture.
There are multiple ways by which PM might cause or be associated with adverse health effects. First, the
reported PM effect estimates in epidemiologic studies may reflect independent PM effects on respiratory
and cardiovascular health. Second, ambient PM may be serving as an indicator of complex ambient air
pollution mixtures that share the same source as PM (i.e., combustion of sulfur-containing fuels or motor
vehicle emissions). Finally, copollutants may mediate the effects of PM or PM may influence the toxicity
of copollutants.
Epidemiologists use the term "interaction" or "effect modification" to denote the departure from
the observed joint risk from what might be expected based on the separate effects of the factors. These
possibilities are not necessarily exclusive. In addition, confounding can result in the production of an
association between adverse health effects and PM that is actually attributable to another factor that is
December 2008
1-15
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
associated with PM in a particular study. Multivariate models are the most widely used strategy to address
confounding in epidemiologic studies, but such models are not always easily interpreted when assessing
effects of covarying pollutants such as 03, S02 and N02.
Inferring causation requires consideration of potential confounders. In confounding, the apparent
effect of the exposure of interest is distorted because the effect of an extraneous factor is mistaken for or
mixed with the actual exposure effect, which may be null. When associations are found in epidemiologic
studies, one approach to remove spurious associations from possible confounders is to control for
characteristics that may differ between exposed and unexposed persons; this is frequently termed
"adjustment." Multivariable regression models constitute one tool for estimating the association between
exposure and outcome after adjusting for characteristics of participants that might confound the results.
The use of multipollutant regression models has been the prevailing approach for controlling potential
confounding by copollutants in air pollution health effects studies. Finding the likely causal pollutant
from multipollutant regression models is made difficult by the possibility that one or more air pollutants
may be acting as a surrogate for an unmeasured or poorly-measured pollutant or for a particular mixture
of pollutants. Further, the correlation between the air pollutant of interest and various copollutants may
show temporal and spatial discongruities that can influence exposures and health effects. Thus, results of
models that attempt to distinguish gaseous and particle effects must be interpreted with caution. Despite
these limitations, the use of multipollutant models is still the prevailing approach employed in most air
pollution epidemiologic studies, and may provide some insight into the potential for confounding or
interaction among pollutants.
Another way to adjust for potential confounding is through stratified analysis, i.e., examining the
association within homogeneous groups with respect to the confounding variable. Stratified analysis can
also be used to examine potential effect modification. The use of stratified analyses has an additional
benefit: it allows examination of effect modification through comparison of the effect estimates across
different groups. If investigators successfully measured characteristics that distort the results, adjustment
of these factors help separate a spurious from a true causal association. Appropriate statistical adjustment
for confounders requires identifying and measuring all reasonably expected confounders. Deciding which
variables to control for in a statistical analysis of the association between exposure and disease depends
on knowledge about possible mechanisms and the distributions of these factors in the population under
study. Identifying these mechanisms makes it possible to control for potential sources that may result in a
spurious association.
Measurement error is another problem encountered when adjusting for spurious associations.
Controlling for confounders, whether by adjustment or stratification, is only successful when the
confounder is well-measured. Considered together, the effects of a well-measured covariate may be
December 2008
1-16
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
overestimated in contrast to a covariate measured with greater error. There are several components that
contribute to exposure measurement error in these studies, including the difference between true and
measured ambient concentrations, the difference between average personal exposure to ambient pollutants
and ambient concentrations at central monitoring sites, and the use of average population exposure rather
than individual exposure estimates. Previous assessments have examined the role of measurement error in
time-series epidemiologic studies using simulated data and mathematical analyses and suggested that
"transfer of effects" would only occur under unusual circumstances (i.e., "true" predictors having high
positive or negative correlation; substantial measurement error; or extremely negatively correlated
measurement errors) (U.S. EPA, 2004).
Confidence that unmeasured confounders are not producing the findings is increased when multiple
studies are conducted in various settings using different subjects or exposures; each of which might
eliminate another source of confounding from consideration. Thus, multicity studies which use a
consistent method to analyze data from across locations with different levels of covariates can provide
insight on potential confounding in associations. The number and degree of diversity of covariates, as
well as their relevance to the potential confounders, remain matters of scientific judgment. Intervention
studies, because of their experimental nature, can be particularly useful in characterizing causation.
In addition to clinical and epidemiologic studies, the tools of experimental biology have been
valuable for developing insights into human physiology and pathology. Laboratory tools have been
extended to explore the effects of putative toxicants on human health, especially through the study of
model systems in other species. Background knowledge of the biological mechanisms by which an
exposure might or might not cause disease can prove crucial in establishing, or negating, a causal claim.
At the same time, species can differ from each other in fundamental aspects of physiology and anatomy
(e.g., metabolism, airway branching, hormonal regulation) that may limit extrapolation. Testable
hypotheses about the causal nature of proposed mechanisms or MOAs are central to utilizing
experimental data in causal determinations.
Interpretations of experimental studies of air pollution effects in animals, as in the case of
environmental comparative toxicology studies, are affected by limitations associated with extrapolation
models. The differences between humans and rodents with regard to pollutant absorption and distribution
profiles based on breathing pattern, exposure dose, and differences in lung structure and anatomy all have
to be taken into consideration. Also, in spite of a high degree of homology and the existence of a high
percentage of orthologous genes across human and rodents (particularly mice), extrapolation of molecular
alterations at the gene level is complicated by species-specific differences in transcriptional regulation.
Given these molecular differences, there are uncertainties associated with quantitative extrapolations at
December 2008
1-17
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
this time between laboratory animals and humans of observed pollutant-induced pathophysiological
alterations under the control of widely varying biochemical, endocrine, and neuronal factors.
1.5.4. Application of Framework for Causal Determination
EPA uses a two-step approach to evaluate the scientific evidence on health or environmental effects
of criteria pollutants. The first step determines the weight of evidence in support of causation and
characterizes the strength of any resulting causal classification. The second step includes further
evaluation of the quantitative evidence regarding the concentration-response relationships and the loads or
levels, duration and pattern of exposures at which effects are observed.
To aid judgment, various "aspects"1 of causality have been discussed by many philosophers and
scientists. The most widely cited aspects of causality in epidemiology, and public health, in general, were
articulated by Sir Austin Bradford Hill in 1965 and have been widely used (CDC, 2004; IARC, 2006;
IOM, 2008; U.S. EPA, 2005a). Several adaptations of the Hill aspects have been used in aiding causality
judgments in the ecological sciences (Adams, 2003; Buck et al., 2000; Collier, 2003; Fox, 1991; Gerritsen
et al., 1998). These aspects (Hill, 1965) have been modified (Table 1-2) for use in causal determinations
specific to health and welfare effects or pollutant exposures.2 Some aspects are more likely than others to
be relevant for evaluating evidence on the health or environmental effects of criteria air pollutants. For
example, the analogy aspect does not always apply, especially for the gaseous criteria pollutants, and
specificity would not be expected for multi-etiologic health outcomes such as asthma or cardiovascular
disease, or ecological effects related to acidification. Aspects that usually play a larger role in
determination of causality are consistency of results across studies, coherence of effects observed in
different study types or disciplines, biological plausibility, exposure-response relationship, and evidence
from "natural" experiments.
1 The "aspects" described by Hill (1965) have become, in the subsequent literature, more commonly described as "criteria." The original term
"aspects" is used here to avoid confusion with 'criteria' as it is used, with different meaning, in the Clean Air Act.
2 The Hill aspects were developed for interpretation of epidemiologic results. They have been modified here for use with a broader array of data,
i.e., epidemiologic, controlled human exposure, and animal toxicological studies, as well as in vitro data, and to be more consistent with EPA's
Guidelines for Carcinogen Risk Assessment.
December 2008
1-18
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Table 1-2. Aspects to aid in judging causality.
Consistency of the observed association. An inference of causality is strengthened when a
pattern of elevated risks is observed across several independent studies. The reproducibility
of findings constitutes one of the strongest arguments for causality. If there are discordant
results among investigations, possible reasons such as differences in exposure, confounding
factors, and the power of the study are considered.
Strength Of the observed association. The finding of large, precise risks increases confidence
that the association is not likely due to chance, bias, or other factors. However, given a truly
causal agent, a small magnitude in the effect could follow from a lower level of exposure, a
lower potency, or the prevalence of other agents causing similar effects. While large effects
support causality, modest effects therefore do not preclude it.
Specificity of the observed association. As originally intended, this refers to increased inference
of causality if one cause is associated with a single effect or disease (Hill, 1965). Based on
our current understanding this is now considered one of the weaker guidelines for causality;
for example, many agents cause respiratory disease and respiratory disease has multiple
causes. At the scale of ecosystems, as in epidemiology, complexity is such that single agents
causing single effects, and single effects following single causes, are extremely unlikely. The
ability to demonstrate specificity under certain conditions remains, however, a powerful
attribute of experimental studies. Thus, although the presence of specificity may support
causality, its absence does not exclude it.
Temporal relationship Of the observed association. Evidence of a temporal sequence between
the introduction of an agent, and appearance of the effect, constitutes another argument in
favor of causality.
Biological gradient (exposure-response relationship). A clear exposure-response relationship
(e.g., increasing effects associated with greater exposure) strongly suggests cause and effect,
especially when such relationships are also observed for duration of exposure (e.g., increasing
effects observed following longer exposure times). There are, however, many possible
reasons that a study may fail to detect an exposure-response relationship. Thus, although the
presence of a biologic gradient may support causality, the absence of an exposure-response
relationship does not exclude a causal relationship.
Biological plausibility. An inference of causality tends to be strengthened by consistency with data
from experimental studies or other sources demonstrating plausible biological mechanisms. A
December 2008
1-19
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
proposed mechanistic linking between an effect, and exposure to the agent, is an important
source of support for causality, especially when data establishing the existence and
functioning of those mechanistic links are available. A lack of biologic understanding,
however, is not a reason to reject causality.
Coherence. An inference of causality from epidemiologic associations may be strengthened by
other lines of evidence (e.g., clinical and animal studies) that support a cause-and-effect
interpretation of the association. Evidence on ecological or welfare effects may be drawn
from a variety of experimental approaches (e.g., greenhouse, laboratory, and field) and
subdisciplines of ecology (e.g., community ecology, biogeochemistry and
paleological/historical reconstructions). The coherence of evidence from various fields
greatly adds to the strength of an inference of causality. The absence of other lines of
evidence, however, is not a reason to reject causality.
Experimental evidence. The strongest evidence for causality can be provided when a change in
exposure brings about a change in occurrence or frequency of health or welfare effects.
Analogy. Structure activity relationships and information on the agent's structural analogs can
provide insight into whether an association is causal. Similarly, information on mode of
action for a chemical, as one of many structural analogs, can inform decisions regarding
likely causality.
While these aspects provide a framework for assessing the evidence, they do not lend themselves to
being considered in terms of simple formulas or fixed rules of evidence leading to conclusions about
causality (Hill, 1965). For example, one cannot simply count the number of studies reporting statistically
significant results or statistically nonsignificant results and reach credible conclusions about the relative
weight of the evidence and the likelihood of causality. Rather, these important considerations are taken
into account with the goal of producing an objective appraisal of the evidence, informed by peer and
public comment and advice, which includes weighing alternative views on controversial issues.
Additionally, it is important to note that the aspects in Table 1-2 cannot be used as a strict checklist, but
rather to determine the weight of the evidence for inferring causality. In particular, not meeting one or
more of the principles does not automatically preclude a determination of causality (e.g., see discussion in
CDC, 2004).
1.5.5. First Step—Determination of Causality
In the ISA, EPA assesses the results of recent relevant publications, building upon evidence
available during the previous NAAQS review, to draw conclusions on the causal relationships between
December 2008
1-20
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
relevant pollutant exposures and health or environmental effects. This ISA uses a five-level hierarchy that
classifies the weight of evidence for causation, not just association1; that is, whether the weight of
scientific evidence makes causation at least as likely as not, in the judgment of the reviewing group. In
developing this hierarchy, EPA has drawn on the work of previous evaluations, most prominently the
IOM's Improving the Presumptive Disability Decision-Making Process for Veterans (IOM, 2008), EPA's
Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), and the U.S. Surgeon General's smoking
reports (CDC, 2004). In the ISA, EPA uses a series of five descriptors to characterize the weight of
evidence for causality. This weight of evidence evaluation is based on various lines of evidence from
across the health and environmental effects disciplines. These separate judgments are integrated into a
qualitative statement about the overall weight of the evidence and causality. The five descriptors for
causal determination are described in Table 1-3.
For PM, this determination of causality step involved a rather complex evaluation of evidence for
different PM indices, different types of health or environmental effects, and for short- and long-term
exposure periods. Determination of causality was made for both the PM measure (PMi0, PMi0.2.5, PM2 5,
and ultrafine particles, to the extent evidence was available for each measure) and for the overall effect
category. In the evaluation of health effects findings in Chapter 6 (for short-term exposure) and Chapter 7
(for long-term exposure), evidence was evaluated for health outcome categories, such as cardiovascular
effects, and then conclusions were drawn based upon the integration of evidence from across disciplines
(e.g., epidemiology, clinical studies and toxicology) and also across the suite of related individual health
outcomes. These chapters initially summarize and evaluate findings for individual health outcomes, then
integrate the results in summary sections to draw conclusions on causality for each PM indicator. In the
integrative synthesis and conclusions in Chapter 2, the ISA presents causality determinations and a
summary of the underlying basis for those determinations for the PM indicator (e.g., PM25), for the
exposure time period (e.g., short- and long-term exposure) and for the major health endpoint categories.
1 It should be noted that the CDC and IOM frameworks use a four-category hierarchy for the strength of the evidence. A five-level hierarchy is
used here to be consistent with the EPA Guidelines for Carcinogen Risk Assessment and to provide a more nuanced set of categories.
December 2008
1-21
DRAFT-DO NOT QUOTE OR CITE
-------
Table 1 -3. Weight of evidence for causal determination.
Health Effects
Ecological and Welfare Effects
Causal Evidence is sufficient to conclude that there is a causal relationship
relationship between relevant pollutant exposures and the health outcome. That
is, a positive association has been observed between the pollutant
and the outcome in studies in which chance, bias, and confounding
could be ruled out with reasonable confidence. Evidence includes, for
example, controlled human exposure studies; or observational studies
that cannot be explained by plausible alternatives or are supported by
other lines of evidence (e.g. animal studies or mode of action
information). Evidence includes replicated and consistent high-quality
studies by multiple investigators.
Evidence is sufficient to conclude that there is a causal relationship
between relevant pollutant exposure and the outcome. Causality is
supported when an association has been observed between the
pollutant and the outcome in studies in which chance, bias, and
confounding could be ruled out with reasonable confidence.
Controlled exposure (laboratory or small- to medium-scale field
studies) provides the strongest evidence for causality, but the scope
of inference may be limited. Generally, determination is based on
multiple studies conducted by multiple research groups, and
evidence that is considered sufficient to infer a causal relationship is
usually obtained from the joint consideration of many lines of
evidence that reinforce each other.
Likely to be a Evidence is sufficient to conclude that a causal relationship is likely to
causal exist between relevant pollutant exposures and health outcome but
relationship important uncertainties remain. That is, a positive association has
been observed between the pollutant and the outcome in studies in
which chance and bias can be ruled out with reasonable confidence
but potential issues remain. For example: a) observational studies
show positive associations but copollutant exposures are difficult to
address and/or other lines of evidence (controlled human exposure,
animal, or mode of action information) are limited or inconsistent; or b)
animal evidence from multiple studies, sex, or species is positive but
limited or no human data are available. Evidence generally includes
replicated and high-quality studies by multiple investigators.
Evidence is sufficient to conclude that there is a likely causal
association between relevant pollutant exposures and the outcome.
That is, an association has been observed between the pollutant
and the outcome in studies in which chance, bias and confounding
are minimized, but uncertainties remain. For example, field studies
show a relationship, but suspected interacting factors cannot be
controlled, and other lines of evidence are limited or inconsistent.
Generally, determination is based on multiple studies in multiple
research groups.
Suggestive of Evidence is suggestive of a causal relationship between relevant
a causal pollutant exposures and the health outcome, but is limited because
relationship chance, bias and confounding cannot be ruled out. For example, at
least one high-quality study shows a positive association but the
results of other studies are inconsistent.
Evidence is suggestive of an association between relevant pollutant
exposures and the outcome, but chance, bias and confounding
cannot be ruled out. For example, at least one high-quality study
shows an association, but the results of other studies are
inconsistent.
Inadequate to Evidence is inadequate to determine that a causal relationship exists
infer a causal between relevant pollutant exposures and health outcome. The
relationship available studies are of insufficient quantity, quality, consistency or
statistical power to permit a conclusion regarding the presence or
absence of an association between relevant pollutant exposure and
the outcome.
The available studies are of insufficient quality, consistency or
statistical power to permit a conclusion regarding the presence or
absence of an association between relevant pollutant exposure and
the outcome.
Suggestive of Evidence is suggestive of no causal relationship between relevant
no causal pollutant exposures and health outcome. Several adequate studies,
relationship covering the full range of levels of exposure that human beings are
known to encounter and considering susceptible or vulnerable
subpopulations, are mutually consistent in not showing a positive
association between exposure and the outcome at any level of
exposure.
Several adequate studies, examining relationships between relevant
exposures and outcomes, are consistent in failing to show an
association between exposure and the outcome at any level of
exposure.
1.5.6. Second Step—Evaluation of Response
1 Beyond judgments regarding causality are questions relevant to quantifying health or
2 environmental risks based on our understanding of the quantitative relationships between pollutant
3 exposures and health or welfare effects.
Effects on Human Populations
4 Important questions regarding quantitative relationships include:
5 ¦ What is the concentration-response or dose-response relationship in the human population?
December 2008
1-22
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
¦ What is the interrelationship between incidence and severity of effect?
¦ What exposure conditions (dose or exposure, duration and pattern) are important?
¦ What subpopulations appear to be differentially affected i.e., more susceptible or vulnerable
to effects?
To address these questions the second step of the EPA framework evaluated the entirety of
policy-relevant quantitative evidence regarding the concentration-response relationships including levels
of pollutant and exposure durations at which effects were observed, and subpopulations that differ from
the general population. This integration of evidence resulted in identification of a study or set of studies
that best approximated the concentration response relationship for the U.S. population, given the current
state of knowledge and the uncertainties that surrounded these estimates.
To accomplish this integration, evidence from multiple and diverse types of studies was considered.
Response was evaluated over a range of observations which was determined by the type of study and
methods of exposure or dose and response measurements. Results from different protocols were
compared and contrasted. Animal data also informed evaluation of concentration-response, particularly
relative to dosimetry, modes of action, and characteristics of susceptible subpopulations. For some health
outcomes, the probability and severity of health effects and associated uncertainties can be characterized.
Chapter 2 presents the integrated findings informative for evaluation of population risks.
An important consideration in characterizing the public health impacts associated with exposure to
a pollutant is whether the concentration-response relationship is linear across the full concentration range
encountered, or if nonlinear relationships exist along any part of this range. Of particular interest is the
shape of the concentration-response curve at and below the level of the current standards. The shape of
the concentration-response curve varies, depending on the type of health outcome, underlying biological
mechanisms and dose. At the human population level, however, various sources of variability and
uncertainty tend to smooth and "linearize" the concentration-response function (such as the low data
density in the lower concentration range, possible influence of measurement error, and individual
differences in susceptibility to air pollution health effects). Additionally, many chemicals and agents may
act by perturbing naturally occurring background processes that lead to disease, which also linearizes
population concentration-response relationships (Clewell and Crump, 2005; Crump et al., 1976; Hoel,
1980). These attributes of population dose-response may explain why the available human data at ambient
concentrations for some environmental pollutants (e.g., PM, ozone, lead [Pb], secondhand tobacco smoke,
radiation) do not exhibit evident thresholds for cancer or noncancer health effects, even though likely
mechanisms include nonlinear processes for some key events. These attributes of human population
December 2008
1-23
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
dose-response relationships have been extensively discussed in the broader epidemiologic literature (e.g.,
Rothman and Greenland, 1998).
Effects on Ecosystems or Public Welfare
Key questions for understanding the quantitative relationships between exposure (or concentration
or deposition) to a pollutant and risk to ecosystems or the public welfare include:
¦ What elements of the ecosystem (e.g., types, regions, taxonomic groups, populations,
functions, etc.) appear to be affected, or are more sensitive to effects?
¦ Under what exposure conditions (amount deposited or concentration, duration and pattern)
are effects seen?
¦ What is the shape of the concentration-response or exposure-response relationship?
Evaluations of causality typically characterize how the probability of ecological and welfare effects
change in response to exposure. A challenge to the quantification of exposure-response relationships for
ecological effects is the variability across ecosystems. Ecological responses are evaluated within the range
of observations, so a quantitative relationship may be determined for a given ecological system and scale.
There is great regional and local variability in ecosystems, thus an exposure-response relationship
generally cannot be determined at the larger national or even regional scale. Quantitative relationships
therefore are available site by site. For example, an ecological response to deposition of a given pollutant
can differ greatly between ecosystems. Where results from greenhouse or animal ecotoxicological studies
are available, they may be used to aid in characterizing exposure-response relations, particularly relative
to mechanisms of action, and characteristics of sensitive biota.
1.5.7. Concepts in Evaluating Adversity of Health Effects
In evaluating the health evidence, a number of factors can be considered in determining the extent
to which health effects are "adverse" for health outcomes such as changes in lung function. What
constitutes an adverse health effect may vary between populations. Some changes in healthy individuals
may not be considered adverse while those of a similar type and magnitude are potentially adverse in
more susceptible individuals.
The American Thoracic Society (ATS) published an official statement titled What Constitutes an
Adverse Health Effect of Air Pollution? (ATS, 2000). This statement updated the guidance for defining
adverse respiratory health effects that had been published 15 years earlier (ATS, 1985), taking into
account new investigative approaches used to identify the effects of air pollution and reflecting concern
December 2008
1-24
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
for impacts of air pollution on specific susceptible groups. In the 2000 update, there was an increased
focus on quality of life measures as indicators of adversity and a more specific consideration of
population risk. Exposure to air pollution that increases the risk of an adverse effect to the entire
population is viewed as adverse, even though it may not increase the risk of any identifiable individual to
an unacceptable level. For example, a population of asthmatics could have a distribution of lung function
such that no identifiable individual has a level associated with significant impairment. Exposure to air
pollution could shift the distribution such that no identifiable individual experiences clinically relevant
effects; this shift toward decreased lung function, however, would be considered adverse because
individuals within the population would have diminished reserve function and, therefore, would be at
increased risk to further environmental insult.
1.6. Summary
This first external review draft ISA is a concise review, synthesis, and evaluation of the most
policy-relevant science, and communicates critical science judgments relevant to the NAAQS review. It
reviews the most edpolicy-relevant evidence from health and environmental effects studies, including
mechanistic evidence from basic biological science. Annexes to the ISA provide additional details of the
literature published since the last review. A framework for making critical judgments concerning causality
was presented in this chapter. It relies on a widely accepted set of principles and standardized language to
express evaluation of the evidence. This approach can bring rigor and clarity to the current and future
assessments. This ISA should assist EPA and others, now and in the future, to accurately represent what is
presently known—and what remains unknown—concerning the effects of PM on human health and
public welfare.
December 2008
1-25
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Chapter 2. Integrative Health Effects
Overview
The subsequent chapters of this ISA will present the most policy-relevant information related to the
review of the NAAQS for PM. This chapter integrates the key findings from the disciplines evaluated in
this current assessment of the PM scientific literature, which includes the atmospheric sciences, ambient
air data analyses, exposure assessment, dosimetry, and health studies (e.g., toxicological, human clinical,
and epidemiologic). The EPA framework for causal determinations described in Chapter 1 has been
applied to the body of evidence in order to judge the scientific data that examines the association between
exposure to PM and health effects in a two-step process. The first step wass to determine the weight of
evidence in support of causation at relevant pollutant exposures and characterize the strength of any
resulting causal classification. The EPA framework applied here employed a five-level hierarchy for
causal determination:
¦ Causal relationship
¦ Likely to be a causal relationship
¦ Suggestive of a causal relationship
¦ Inadequate to infer a causal relationship
¦ Suggestive of no causal relationship
The second step evaluated the entirety of policy-relevant quantitative evidence regarding the
concentration-response relationships including levels and exposure durations at which effects were
observed, and subpopulations that were more susceptible or vulnerable to PM exposure than the general
population. This integration of evidence resulted in identification of a study or set of studies that best
estimated the concentration-response relationships for the U.S. population, given the current state of
knowledge. Together the two steps in the framework led to: (1) causal determinations for a range of health
outcomes, and (2) characterization of the magnitude of these responses, including susceptible or
vulnerable subpopulations, over a range of ambient concentrations.
This chapter summarizes and integrates the newly available scientific evidence that best informs
consideration of the policy-relevant questions that frame this assessment, presented in Chapter 1.
Section 2.1 discusses the trends in ambient concentrations and sources of PM and provides a brief
summary of ambient air quality for short- and long-term exposure durations. Section 2.2 presents the
evidence regarding personal exposure to ambient PM in outdoor and indoor microenvironments, and it
December 2008
2-1
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
discusses the relationship between ambient PM concentrations and exposure to PM from ambient sources.
Section 2.3 integrates the evidence for studies that examine the development of health effects in response
to short- and long-term exposure to PM and discusses important uncertainties identified in the
interpretation of the scientific evidence. In addition, the section discusses the evidence from recent studies
that examined the association between PM components or sources, instead of mass and health effects.
Finally, Section 2.4 presents the public health impacts associated with exposure to PM, which includes
evidence for potentially susceptible and vulnerable populations to PM exposure.
2.1. Concentrations and Sources of Atmospheric PM
2.1.1. Ambient PM Variability and Correlations
Advances in understanding the spatiotemporal distribution of PM mass and constituents have
recently been made, particularly with regard to PM2 5 mass and chemical composition and ultrafine
concentrations. Emphasis in this ISA was on the period from 2005-2007 so that the most recent validated
EPA Air Quality System (AQS) data were used. Note, however, that a majority of U.S. counties were not
represented in AQS data, since their population densities fell below the regulatory monitoring threshold
for PM. Moreover, monitors reporting to AQS were not uniformly distributed across the U.S. or within
counties, and conclusions drawn from AQS data may not apply equally to all parts of a geographic region.
Furthermore, biases can exist for some PM constituents (and hence total mass) owing to volatilization
losses of nitrates and other semi-volatile compounds, and, conversely, to retention of particle-bound water
by hygroscopic species. The degree of spatial variability in PM was likely to be region-specific and
strongly influenced by region-specific sources and meteorological and topographic conditions.
2.1.1.1. Spatial Variability across the U.S.
County-scale, 24-h average concentration data for PMi0 and PM2 5 for 2005-2007 showed
considerable variability across the U.S. Figures 3-6 and 3-7 show county-scale coverage and average
concentrations for PM10 and PM25. For PM10, the highest reported annual average concentrations
(>51 |ag/m3) occurred in two counties in southern California and five counties in southern Arizona and
central New Mexico. The lowest reported annual average PM10 concentrations (< 20 (.ig/nr1) were within
114 counties distributed fairly uniformly across the U.S. For PM25, the highest reported annual average
concentration (>20 (.ig/nr1) were reported for six counties within the San Joaquin Valley and inland
southern California, as well as Jefferson County, AL (containing Birmingham) and Allegheny County, PA
December 2008
2-2
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
(containing Pittsburgh). The lowest reported annual average PM2 5 concentrations (<12 (.ig/ni3) were
contained within 237 counties distributed throughout the west, northeast, Florida and the Carolinas.
The concentration of PM25 relative to that of PM10 varied substantially by location, with a larger
fraction of PM mass in the coarse mode in cities with dryer climates (e.g., Phoenix and Denver) and a
larger fraction in the fine mode in eastern U.S. cities (e.g., Pittsburgh and Philadelphia). Limiting the
differential calculation of PM10-2 5 to low volume federal reference method (FRM) PMi0 and PM25
monitors helps reduce sampling artifacts resulting from subtracting two independent mass measurements.
However, this results in poor geographic coverage since few sites have the appropriate co-located
monitors for computing this difference. Figure 3-8 contains all U.S. counties where co-located low
volume FRM data was available for this calculation. Although the general understanding of PM
differential settling leads to an expectation of greater spatial heterogeneity in the PMi0_2.5 fraction,
deposition of particles as a function of size depends strongly on local meteorological conditions. Current
data coverage is insufficient to draw any meaningful conclusions regarding the spatial distribution of
PMio-2.5-
Spatial variability in PM2 5 components obtained from the Chemical Speciation Network (CSN)
varied considerably by species, including organic carbon (OC), elemental carbon (EC), sulfate (S042 ).
nitrate (N03 ) and ammonium (NH4) (see Section 3.5.1.1). The highest annual average OC
concentrations (>5 |_ig/m3) were observed in the western and southeastern U.S. Concentrations in the
western U.S. peaked in the fall and winter, while concentrations in the Southeast peaked anytime between
spring and fall. EC exhibited less seasonality than OC and was particularly stable in the eastern half of the
U.S. Annual average EC concentrations greater than 1.5 (ig/m3 were present in Los Angeles, Pittsburgh,
New York and El Paso. Concentrations of S042 were higher in the eastern U.S. as a result of higher S02
emissions in the East, compared with the West. There is also considerable seasonal variability with higher
S042 concentrations in the summer months when the oxidation of S02 proceeds at a faster rate than
during the winter. N03 concentrations were highest in California, with annual averages >4 |_ig/m3 at
many monitoring locations. There were also elevated levels of N03 in the Upper Midwest (>2 (.ig/m3).
particularly in the winter. In general, N03 was higher in the winter across the country, in part as a result
of temperature-driven partitioning and volatilization. Exceptions existed in Los Angeles and Riverside,
where high N03 readings appeared year round. Concentrations of NH44" were similar to concentrations of
N03 or S042 throughout the U.S. Clearly, there is variation in both PM2 5 mass and composition by city,
some of which might be due to regional differences; however, there are too many controlling variables
(e.g. meteorology, sources, topography) which are too poorly characterized at this scale to allow
conclusions to be drawn regarding PM2 5 composition across all cities within a given geographic region.
December 2008
2-3
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Variability in PM2 5 components across the U.S. was examined by focusing on fifteen metropolitan
areas chosen based on their geographic distribution and coverage in recent health effects studies (see
Section 3.5.1.1). The urban areas selected were Atlanta, Birmingham, Boston, Chicago, Denver, Detroit,
Houston, Los Angeles, New York, Philadelphia, Phoenix, Pittsburgh, Riverside, Seattle and St. Louis. On
an annual average basis, sulfate was the dominant PM2 5 component in the eastern cities, ranging from
42% of PM2 5 mass in Chicago to 56% in Pittsburgh. Organic carbon mass (OCM) was the next largest
component. In the western cities, OCM was the largest constituent of PM2.5 on an annual basis, ranging
from 34% in Los Angeles to 58% in Seattle. Sulfate, nitrate and crustal material were all important
components in the western cities analyzed. Sulfate ranged from 18% in Denver to 32% in Los Angeles.
Nitrate was particularly large in Riverside (22%), Los Angeles (19%) and Denver (15%); crustal material
constituted a substantial fraction of PM2 5 year-round in Phoenix (28%) and Denver (16%), and during the
summer in Houston (26%), even though the annual average was much lower (11%).
2.1.1.2. Spatial Variability on the Urban and Neighborhood Scales
In general, PMi0 has a shorter atmospheric lifetime than PM2 5 because PMi0 contains larger
particles which have higher settling velocity. As a result, local emission sources often dominate PMi0
annual average mass concentrations at particular monitors, while PM2 5 mass concentrations are more
homogeneously distributed (see Section 3.5.1.2). Therefore, as an example, using the 15 cities listed
above, there was considerably less decline in the correlation between monitors as a function of distance
for PM2 5 than for PM10. Furthermore, correlations between PM2 5 concentrations exhibited substantially
less scatter. For PM10, Atlanta, Boston, Denver, Los Angeles, New York City, Philadelphia, Phoenix,
Pittsburgh and Riverside all showed relatively high correlations as a function of distance (average
correlation of 0.75 at 40 km or greater monitor separation), while Birmingham, Chicago, Detroit, Houston
and St. Louis had correlations that dropped off much more quickly with distance (average correlation of
0.75 at 6 km or less monitor separation). The Seattle data only included two PMi0 monitoring sites, thus
providing insufficient information to draw any conclusions. For PM2 5, most metropolitan areas exhibited
high correlations (generally >0.75) out to a distance of 100 km. Notable exceptions were Denver, Los
Angeles and Riverside, where correlations dropped below 0.75 somewhere between 20 and 50 km.
Insufficient data were available in the 15 metropolitan areas to perform similar analyses for PMi0.2.5 using
co-located, low volume FRM monitors.
Population density and associated building density are important determinants of the spatial
distribution of PM concentrations. Inter-sampler correlations as a function of distance between monitors
obtained for sampler pairs located less than 4 km apart (i.e., on a neighborhood scale) showed a shallower
December 2008
2-4
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
slope for PM2 5 than for PM10. Average correlation was maintained at 0.93 for PM2 5, while it dropped to
0.70 for PM10 (see Section 3.5.1.3).
Few studies performed direct comparisons of ultrafine particle measurements at multiple locations
within an urban area. A decrease in the number of ultrafine particles was demonstrated with shifts from a
dominant mode at around 10 nm within 20 m of a freeway to a flattened dominant mode at around 50 nm
at a distance of roughly 100-150 m. At the same time, accumulation mode particle number concentration
remained relatively constant to within -300 m from the freeway. These findings suggest a high degree of
spatial heterogeneity in ultrafine particles compared with accumulation mode particles on the urban scale.
2.1.2. Temporal Variability
Trends in PMi0 concentrations show a steady decline from 1988 to 2007 in all 10 EPA Regions. A
steady decrease in PM2 5 concentrations from 1999 (the beginning of nationwide monitoring for PM2 5) to
2007 was observed in all 10 EPA Regions, with the three-year average of the 98th percentile of 24-h
PM2 5 concentrations dropping 10% over this time period.
Using hourly PM observations in the 15 metropolitan areas, diel variation showed peaks that differ
by pollutant and region. For PM10, all areas showed a gradual morning increase in mean concentrations
starting at approximately 6:00 am on weekdays, corresponding with both the start of morning rush hour
and break-up of the overnight inversion layer. The magnitude and duration of this peak varied
considerably by metropolitan area. For PM2 5, a similar morning peak was observed starting at
approximately 6:00 am in all cities except Pittsburgh, where elevated overnight PM2 5 obscures any
morning peak. There was also an evening PM2 5 concentration peak that was broader than the morning
peak and extended into the overnight period, reflecting the concentration increase caused by the usual
collapse of the mixed layer after sundown (see Section 3.5.2.3).
Studies indicate that ultrafine particles in urban environments exhibit similar two-peaked diel
patterns in Los Angeles and the San Joaquin Valley as well as in Kawasaki City, Japan and Copenhagen,
Denmark (see Section 3.5.2.3). The afternoon peak in ultrafine particles likely represents the combination
of primary source emissions such as evening rush hour traffic and photochemical formation of secondary
organic aerosol.
2.1.3. Correlations between Copollutants
Correlations between PM and gaseous copollutants including S02, N02, carbon monoxide (CO)
and 03 varied both seasonally and spatially between and within metropolitan areas. On average, PM10 and
December 2008
2-5
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
PM2 5 were correlated with each other better than with the gaseous copollutants. There was relatively little
seasonal variability in the mean correlation between PM in both size fractions and S02 and N02. CO,
however, showed higher correlations with PM10 and PM2 5 on average in the winter compared with the
other seasons. This seasonality results in part because a larger fraction of PM is primary in origin during
the winter. To the extent that this primary component of PM is associated with common sources of N02
and CO, then higher correlations with these gaseous co-pollutants are to be expected. Increased
atmospheric stability in colder months would also reinforce these associations.
The correlation between daily maximum 8-h average 03 and PM showed the highest degree of
seasonal variability with positive correlations on average in the spring, summer and fall, and negative
correlations on average in the winter. This situation arises as the result of seasonal differences in sources
and photochemical production of secondary PM2 5 and 03. However, this relationship is not found in all
cities examined (e.g., Birmingham, Boston and St. Louis).
2.1.4. Measurement Techniques
Reliable methods have been developed to measure real-time PM mass concentrations (e.g., FDMS-
TEOM). Real-time (or continuous and semi-continuous) measurement techniques are also available for
PM species, such as PILS for multiple ions analysis and AMS for multiple components analysis.
Advances have also been achieved in PM organic speciation (e.g. TD-GC-MS) (For additional
information see Section 3.4).
2.1.5. PM Source Characteristics
PM in the atmosphere contains both primary (i.e., emitted directly by sources) and secondary
components, which can be anthropogenic or natural in origin. Secondary components are produced by the
oxidation of precursor gases such as S02, NOx and ammonia (NH3) and organic compounds. The largest
sources of primary PM2 5 on a nationwide basis are wildfires, road dust, and electricity-generating units
(EGUs), with road dust being the largest single source of PMi0 according to the National Emissions
Inventory (NEI).
Developments in the chemistry of formation of secondary organic aerosol (SOA) indicate that
oligomers are likely a major component of OC in aerosol samples. Until a few years ago, the oxidation of
terpenes and aromatic compounds were considered as sources of SOA, but not the oxidation of isoprene.
However, recent observations suggest that small, but important quantities of SOA are formed from
isoprene oxidation. Gasoline engines have been found to emit a mix of nucleation-mode heavy and large
December 2008
2-6
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
polycyclic aromatic hydrocarbons on which unspent fuel and trace metals condense, while diesel particles
are composed of a soot nucleus on which S042 and hydrocarbons condense. Current inventories of
emissions from combustion sources overestimate the primary component of organic aerosol and
underestimate the semi-volatile components in the emissions. This situation results from the lack of
capture of evaporated semi-volatile components upon dilution in standard emissions tests. As a result,
near-traffic sources of organic aerosol are underestimated, however, farther downwind the overall
formation rate of SOA increases as a result of the oxidation of these semi-volatile components.
2.1.6. Source Contributions to PM
Results of receptor modeling calculations indicate that PM2 5 is produced mainly by combustion of
fossil fuel, either by stationary sources or by transportation. It is apparent that a relatively small number
of source categories, compared to the total number of chemical species that typically are measured in
ambient monitoring source receptor model studies, are needed to account for the majority of the observed
mass of PM in these studies. A compilation of study results shows that secondary sulfate (mainly from
EGUs), nitrate (from the oxidation of NO emitted mainly from transportation and EGUs), and primary
mobile source categories constitute most of PM2 5 (and PM10) in the East. Fugitive dust, found mainly in
the PM10_2 5 size range, represents the largest source of ambient PM10 in many locations in the western
U.S. Quoted uncertainties in the source apportionment of constituents in ambient aerosol samples
typically range from 10 to 50%. An intercomparison of source apportionment techniques indicated that
the same major source categories of PM2 5 were consistently identified by several independent groups
working with the same data sets. Soil-, sulfate-, residual oil-, and salt-associated mass were most clearly
identified by the groups. Other sources with more ambiguous signatures, such as vegetative burning and
traffic-related emissions were less consistently identified.
Spatial variability in source contributions across urban areas is an important consideration in
assessing the likelihood of exposure error in epidemiologic studies relating health endpoints to sources.
Concepts similar to those for using ambient concentrations as surrogates for personal exposures apply
here. Studies for PM2 5 indicate that intra-urban variability increases in the following order: regional (e.g.
secondary S042 from EGUs) < area (e.g. on road mobile sources) < point (e.g. stacks) sources. Only one
study was available for PM10-2.5, indicating a similar ordering, but without a regional component (resulting
from the short lifetime of PMi 0-2.5 compared to transport times on the regional scale).
December 2008
2-7
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
2.1.7. Policy-Relevant Background
The background concentration of PM that is useful for risk and policy assessments informing
decisions about the NAAQS are referred to as policy-relevant background (PRB) concentrations. PRB
concentrations are those concentrations that would occur in the U.S. in the absence of anthropogenic
emissions in continental North America (defined here at the U.S., Canada and Mexico). PRB
concentrations include contributions from natural sources everywhere in the world and from
anthropogenic sources outside these three countries. Background levels so defined facilitate separation of
pollution levels that can be controlled by U.S. regulations (or through international agreements with
neighboring countries) from levels that are generally uncontrollable by the U.S. Contributions to policy-
relevant background (PRB) levels of PM include both primary and secondary natural and anthropogenic
components (see Section 3.6). PRB concentrations for the continental U.S. were estimated using a
deterministic, continental scale chemistry-transport model (CTM) using results from the GEOS-Chem
global scale CTM as boundary conditions. PRB concentrations of PM2 5 were estimated to be less than 1
(ig/m3 on an annual basis and maximum daily average values generally range from 3.1 to 20 |_ig/m3 with a
peak as high as 63 (ig/m3 at the nine national park sites across the U.S. that were used for model
evaluation.
2.2. Human Exposure
This section summarizes the findings from the recent exposure assessment literature, which include
the assessment of exposure to ambient PM, infiltration of ambient PM to indoor environments, and source
apportionment of exposure. This summary is intended to support the interpretation of the findings from
epidemiologic studies. For a more detailed explanation see Section 3.7.
2.2.1. Outdoor Exposure to Ambient PM
The correlation between the PM concentration measured at a central community ambient monitor
and the true community average concentration depends on the spatial distribution of the PM, selection of
the monitoring site chosen to represent the community average, and division of the community by terrain
features or source locations into several sub-communities that differ in the temporal pattern of pollution.
Some studies, conducted mainly in Europe, have found that personal PM2.5 and PMi0 exposures for
pedestrians in street canyons could be much higher than ambient concentrations measured by urban
background ambient monitors. As a result, ambient monitors located at background, central urban, road
December 2008
2-8
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
side, or near-residential sites might not reflect peak exposures to some individuals in a community.
Ambient monitor height also affects estimates of exposure because PM concentration varies as a function
of height. Within a street canyon, changes in wind direction and speed cause significant variability over a
small distance, with findings showing up to a two order of magnitude change in benzo[a]pyrene
concentrations across a street canyon. Wind tunnel studies have shown street canyon effects exist for
suburban and not just for downtown, heavily urbanized settings.
2.2.2. Indoor and Personal Exposure to Ambient PM
PM infiltration factors, Finf, depend on particle size, chemical composition, season, and region of
the country. Infiltration can best be modeled dynamically based on a distribution of air exchange and
deposition PM loss rates rather than being represented by a single value. There is significant variability
within and across regions of the country with respect to indoor exposures to ambient PM. Infiltrated
ambient PM concentrations depend in part on the ventilation properties of the building or vehicle in which
the person is exposed. Season is important to PM infiltration because it affects the ventilation practices
used, and ambient temperature and humidity conditions affect the transport, dispersion, and size
distribution of PM. Residential air exchange rates have been observed to be higher in summer for regions
with low air conditioning usage, and regional differences in air exchange rates (Southwest < Southeast
< Northeast < Northwest) also reflect ventilation practices. Differential infiltration occurs as a function of
PM size and composition. PM infiltration is largest for accumulation mode particles, and decreases for
ultrafine PM lost to diffusion and for coarse particles lost through inertial impaction mechanisms.
Infiltration is also affected by variations in particle composition and volatility. For example, EC or black
carbon (BC) infiltrates more readily than OC. Differential infiltration can affect both exposure estimates
and PM toxicity.
Emission inventories and source apportionment studies suggest that sources of PM exposure vary
by region. Comparison of studies performed in the eastern U.S. with studies performed in the western
U.S. suggest that the contribution of S042 to personal exposure is higher for the East (16-46%) compared
with the West (-4%) and that motor vehicle emissions and secondary N03 are larger sources of personal
exposure for the West (-9%) as compared with the East (-4%). Results of source apportionment studies
of personal exposure to S042 indicate that personal S042 exposures are mainly attributable to ambient
sources. Source apportionment for OC and EC is difficult because they originate from both indoor and
outdoor sources. Exposure to OC of indoor and outdoor origin can be distinguished by the presence of
aliphatic C-H groups generated indoors, since outdoor concentrations of aliphatic C-H are low. Trace
metal studies have shown variable results regarding personal exposure to ambient constituents with
December 2008
2-9
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
significant variation among cities and over seasons that can be related to incinerator operation, fossil fuel
combustion, biomass combustion (wildfires), and presence of crustal materials in the built environment,
among other sources.
2.2.3. Implications for Epidemiologic Studies
Variations in PM and its components could lead to errors in using ambient PM measures as
surrogates for exposures to PM. PM2 5 and PMi0 concentrations are relatively well-correlated across
monitors in the urban areas examined. Correlation coefficients tend to be lower, and concentration
differences tend to be higher between PMi0 monitoring sites than between PM2 5 monitoring sites. Even if
PM2 5 and PMi0 concentrations measured at sites within an urban area are highly correlated, significant
differences in their concentrations can occur on any given day. The degree of spatial uniformity in PMi0
and PM2 5 concentrations in urban areas varies across the country. Current information suggests that
PMio-2.5 and some PM components are more spatially variable than PM2 5. These factors should be
considered in using data obtained from monitoring networks to estimate community-scale human
exposure to ambient PM, and caution should be exercised in extrapolating conclusions obtained from one
urban area to another.
Community, time-series epidemiologic studies use the average community PM concentration as a
surrogate for the average personal exposure to ambient PM. The resulting health effect risk estimate,
based on the average community ambient concentration, differs from the risk that would be estimated if
the average community ambient exposure were used in the epidemiologic study. This difference is given
by the average ambient exposure factor. However, the risk estimate based on the ambient concentration
gives the change in health effects resulting from a change in ambient concentration of PM and is,
therefore, an appropriate measure for risk assessment and risk management. Variations in ambient
concentrations across a community, variations in individual ambient exposures around the community
average, and seasonal or daily variation in the ambient exposure factor may increase standard errors of
PM health effects estimates, making it more difficult to detect a true underlying association between the
correct exposure metric and the health outcome studied. The use of the community average ambient PM
concentration as a surrogate for the community average personal exposure to ambient PM is not expected
to change the principal conclusions from PM epidemiologic studies that use community average health
and pollution data (U.S. EPA, 2004). Several recent studies support this by showing how the ambient
component of personal exposure to PM2 5 could be estimated using various tracer and source
apportionment techniques and that is highly correlated with ambient concentrations of PM2 5. These
studies also show that the non-ambient component of personal exposure to PM2 5 is basically uncorrelated
December 2008
2-10
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
with ambient PM2 5 concentrations. For long-term studies that use differences in long-term community
average ambient PM concentrations as an exposure metric, the effect of possible community-to-
community differences in the average ambient exposure factor or in the average non-ambient exposure
are less understood. For panel epidemiologic studies, the most appropriate exposure metric may depend
on the health outcome measured. However, sufficient information should be obtained to enable
determining the association of the health outcome with ambient concentration, ambient exposure, non-
ambient exposure, and total personal exposure.
A number of studies have examined whether gaseous copollutants could act as surrogates for
exposure to ambient PM. Several studies have concluded that ambient concentrations of 03, N02, and
S02 are associated with the ambient component of personal exposure to total PM2 5 as opposed to the
ambient component of personal exposures to the gases. However, in some studies this result may have
arisen in part because personal exposure to the gases was often beneath the detection limits of the
personal monitoring devices. Thus, the evidence that ambient gases can be considered surrogates of PM2 5
exposure is mixed. It is likely that associations between ambient gases and personal exposure to PM2 5 of
ambient origin exist, but they are complex and vary by season and location.
2.3. Health Effects
This section evaluates the evidence from toxicological, human clinical, and epidemiologic studies
that examined the health effects associated with short- and long-term exposure to PM (i.e., PMi0, PM2 5,
PMio_25and ultrafine particles [0.01-0.1 |_im |). Within this section a discussion of the causal
determinations will be presented by PM size fraction and exposure type (i.e., short- or long-term
exposure) for only those health endpoints in which sufficient evidence was available to conclude that a
specific PM size fraction causes or likely Causes a health effect (i.e., cardiovascular morbidity,
respiratory morbidity, mortality). Although an extensive amount of research has been conducted to
examine PM-related health effects, a limited body of evidence is currently available to examine the
presence or absence of associations between some health outcomes and PM size fractions. Thus, based on
currently available evidence, it is not possible to causally link exposure duration, PM size fraction and
health outcome for all combinations evaluated in this ISA. The evaluation of the aforementioned factors
together has resulted in evidence that is Suggestive of a causal relationship for mortality and respiratory
morbidity in response to short-term exposure to PM2 5, mortality, cardiovascular morbidity and
reproductive and developmental effects in response to long-term exposure to PMi0, and reproductive and
developmental effects in response to long-term exposure to PM2 5. In addition, inadequate evidence to
December 2008
2-11
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
infer a causal relationship exists for all other health outcomes (due to both short- and long-term exposure)
to PM10_2.5 and ultrafine PM. A detailed discussion of the underlying evidence used to formulate each
causal determination can be found in Chapters 6 and 7.
2.3.1. Exposure to PM10
2.3.1.1. Effects of Short-Term Exposure to PM10
Size Fraction
Outcome
Causality Determination
PM10
Cardiovascular morbidity
Likely to be causal
Respiratory morbidity
Likely to be causal
Mortality
Likely to be causal
Cardiovascular Morbidity
The majority of recent evidence for an association between short-term exposure to PMi0 and
cardiovascular (CV) health effects is derived from epidemiologic studies of hospital admissions (HAs)
and emergency room (ED) visits (see Section 6.2.10). Although some regional heterogeneity is evident in
the single-city effect estimates, consistent increases in HAs and ED visits for cardiovascular diseases
(CVD), has been observed across studies, with the majority of estimates ranging from 0.5-1.0% per 10
(ig/m3 increase in PM10 (see Figure 6-1). A detailed examination of specific CV health outcomes has
suggested that ischemic heart disease (IHD) and chronic heart failure (CHF) are responsible for the
majority of PM-related CVD HAs rather than cerebrovascular diseases; however, one large multicity
U.S.-based study provides evidence of an association between PM10 and ischemic stroke. Overall, the new
literature provides consistent evidence for associations between short-term exposure to PMi0 and
increased risk of cardiovascular HAs and ED visits in cities with mean 24-h average concentrations
ranging from 16.8 to 48 (ig/m3.
Human clinical studies which evaluate the effect of PMi0 on measures of cardiovascular function
have not been conducted; however, a few recent animal toxicological studies demonstrated impacts on the
cardiovascular system. A new inhalation study in animals found lowered cardiac contractility upon
exposure to PMi0, while several intratracheal instillation studies found altered vasoreactivity and elevated
levels of systemic inflammatory and blood coagulation markers (see Sections 6.2.1. through 6.2.9). In
addition, several epidemiologic studies have observed physiologic alterations in CV function including:
December 2008 2-12 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
heart rate variability (HRV), systemic markers of inflammation, coagulation, and oxidative stress in cities
with mean 24-h average concentrations ranging from 10.5 to 46.1 |ig/nr\ These findings, along with those
reported in the toxicological literature contribute to the biological plausibility of PM10-related
cardiovascular effects.
Overall, consistent and coherent evidence exists across recent toxicological and epidemiologic
studies, which supports the conclusion that short-term exposure to PMi0 is associated with an increased
risk of cardiovascular morbidity. Furthermore, findings of altered autonomic function, cardiac
contractility, systemic inflammation, coagulation, and vasoreactivity provide biological plausibility that
exposure to PMi0 could lead to more severe effects, including HAs or ED visits for IHD, CHF, or
ischemic stroke. Collectively, the studies evaluated provide sufficient evidence to conclude that a causal
relationship is likely to exist between short-term exposure to ambient concentrations of PM10 and
cardiovascular morbidity.
Respiratory Morbidity
Epidemiologic studies that examined the association between short-term exposure to PMi0 and
respiratory morbidity found consistent positive effects in asthmatic children and adults, but no evidence
of an association in healthy individuals. The majority of the studies that examined the association between
PMio and respiratory symptoms and medication use found an increased risk ranging from -1.0 to 1.75 for
cough, phlegm, difficulty breathing, and bronchodilator use in asthmatic children in cities with mean 24-h
average concentrations ranging from 16.8 to 64.5 (ig/m3. Positive, but less consistent effects for
respiratory symptoms and medication use were observed in asthmatic adults (see Figure 6-7). One study
in the new epidemiologic literature examined the effects of PM10 on pulmonary inflammation, and
observed an association between PM10 and exhaled nitrogen oxide (eNO). An evaluation of respiratory
ED visit and HA studies found consistent positive associations at ambient PM10 concentrations ranging
from 13.3 to 60.8 (ig/m3 (see Section 6.3.8.), among asthmatic children (~ 2% increase) and older adults
with COPD (~ 0 to 3% increase). Although no toxicological or human clinical studies in the new body of
literature examined the effect of short-term exposure to PMi0 on respiratory morbidity, the consistent
epidemiologic evidence alone is sufficient to conclude that a Causal relationship is likely to exist
between short-term exposure to ambient concentrations of PM10 and respiratory morbidity
Mortality
The epidemiologic literature indicates consistent positive associations between short-term exposure
to PMio and all-cause mortality. The multicity studies evaluated reported an approximate 0.12-0.81%
increase in all-cause mortality per 10 (ig/m3 increase in PMi0with 24-h average PMi0 concentrations
December 2008
2-13
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
ranging from 13 to 53.2 (ig/m3 (see Section 6.5). Although respiratory and cardiovascular-related
mortality also show consistent positive effects, only a few multicity studies conducted cause-specific
mortality analyses. Heterogeneity in PM10 mortality risk estimates was observed between cities and
studies, which could be attributed to the lag, averaging time, number of cities and/or copollutants included
in the regression models. An evaluation of the lag structures used in the multicity studies found that the
greatest effects were observed using the previous day's PMi0 concentration (lag 1) or the average of the
same day's and previous day's concentrations (lag 0-1). In addition, the use of a distributed lag model
resulted in slightly larger (by -30%) estimates compared to single-day lags. Regional heterogeneity and
seasonal patterns in PMi0 risk estimates were also observed, with the greatest effects occurring in the
Eastern U.S. and during the summer and transition seasons, spring and fall, respectively. An examination
of potential confounders (i.e., temperature and copollutants) using different study designs (i.e., time series
and case crossover) observed that neither is likely to account for differences in PMio-mortality risk
estimates between studies. However, one Canadian-based multicity study did observe a reduction in the
PMio mortality risk estimate upon the inclusion of N02 in the model. Overall, the consistent evidence
found across epidemiologic studies is sufficient to conclude that a Causal relationship is likely to exist
between short-term exposure to ambient concentrations of PM10 and mortality.
2.3.1.2. Effects of Long-Term Exposure to PM10
Size Fraction Outcome Causality Determination
PM10 Cardiovascular morbidity Suggestive
Respiratory Morbidity
Likely to be causal
Mortality
Suggestive
Reproductive and Developmental
Suggestive
Cancer Inadequate
Respiratory Morbidity
The recent epidemiologic literature focuses on prospective cohort studies, which found consistent
positive associations between long-term exposure to PM10 and respiratory morbidity (see Section 7.3).
U.S.- and European-based multi- and single-city studies have observed an increase in respiratory
symptoms (e.g., bronchitis and cough) in children at annual average concentrations ranging from 7.0 to
34.8 (ig/m3. The strength of the observed association is increased through the results of a cohort
December 2008
2-14
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
consisting of school children in Switzerland, which found a reduction in respiratory symptoms
(i.e., chronic cough, bronchitis, common cold, nocturnal dry cough, and conjunctivitis symptoms) that
coincided with a reduction in ambient PM10 levels. Additional epidemiologic studies conducted in both
the U.S. and abroad also found a relationship between ambient PMi0 levels and decrements in lung
function and lung function growth (Figure 7-4). Although the epidemiologic evidence supports an
increase in respiratory health effects in response to long-term exposure to PMi0, a high correlation
between PMi0 and other pollutants has also been observed, which could potentially confound the PMi0-
respiratory morbidity relationship. Overall, the consistent associations observed across studies and
locations provide sufficient evidence to conclude that a causal relationship is likely to exist between
long-term exposure to ambient concentrations of PM10 and respiratory morbidity
2.3.2. Exposure to PM2.5
2.3.2.1. Effects of Short-Term Exposure to PM2.5
Size Fraction
Outcome
Causality Determination
PM2.5
Cardiovascular morbidity
Causal
Respiratory morbidity
Likely to be causal
Mortality
Likely to be causal
Cardiovascular Morbidity
The large body of evidence from studies that examined the effect of short-term exposure to PM2 5
on cardiovascular morbidity found consistent cardiovascular health effects across epidemiologic, human
clinical and toxicological studies. Epidemiologic studies that examined the effect of PM25 on
cardiovascular ED visits and HAs reported consistent positive associations, with the majority ranging
from ~ 0.5 to 3.4%, for a 10 (ig/m3 increase in PM25 in cities with mean 24-h average concentrations of
13.8-18.8 |_ig/m3 (see Section 6.2.10). The largest U.S.-based multicity study, Medicare Air Pollution
Study (MCAPS), also observed regional heterogeneity (i.e., the largest excess risks occurred in the
Northeast [1.08%]) and seasonal variation (i.e., the largest excess risks occurred during the winter season
[1.49%]) in PM2 5 risk estimates. The PM2 5 ED visit and HA effects observed appear to be driven by IHD
and CHF rather than cerebrovascular diseases. Additional epidemiologic studies that examined
physiologic alterations in cardiovascular function observed changes in HRV, electrocardiogram (ECG)
December 2008
2-15
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
abnormalities, vasomotor function, systemic inflammation, coagulation and oxidative stress at mean 24-h
average concentrations ranging from 7.7-23.7 (ig/m3, which provides biological plausibility for the
development of cardiovascular health effects at ambient PM2 5 concentrations.
Controlled human exposure studies have consistently demonstrated changes in various measures of
cardiovascular function following exposure to PM25. The majority of the new studies described have been
conducted using diesel exhaust (DE) or concentrated ambient particles (CAPs), and provide strong
evidence of PM2 5-induced decreases in HRV and vasomotor function, as well as increases in markers of
systemic oxidative stress (see Section 6.2). An additional study observed a decrease in ST-segment
depression following exposure to DE in a group of older adults with prior myocardial infarction (MI).
Although not consistently observed across studies, some investigators have reported PM2 5-induced
changes in blood pressure (BP), blood coagulation markers, and markers of systemic inflammation (see
Section 6.2).
Additional new toxicological studies have demonstrated that short-term exposure to PM2 5 can
result in cardiovascular health effects. Consistent with evidence from human clinical studies, the most
significant contributions from the current toxicological literature for acute PM2 5-induced cardiovascular
effects are decreased myocardial blood flow following ischemia, changes in vascular reactivity, and
increased cardiac oxidative stress (see Section 6.2). The results for additional cardiovascular health effects
such as HRV, arrhythmia, systemic inflammation, and blood coagulation are mixed, while very few or
weakly designed studies were evaluated for BP and cardiac contractility. Taken together, the results from
the new human clinical and toxicological studies provide coherence and support the biological plausibility
of an association between short-term exposure to PM2 5 and cardiovascular morbidity.
Overall, the consistent and coherent results from epidemiologic, human clinical, and toxicological
studies provides sufficient evidence to conclude that a causal relationship exists between short-term
exposure to ambient concentrations of PM2.5 and cardiovascular morbidity
Respiratory Morbidity
The recent epidemiologic literature that examined the association between short-term exposure to
PM2 5 and respiratory morbidity focused on both respiratory symptoms, which includes medication use,
along with respiratory-related HAs and ED visits. The majority of the studies that examined the
association between PM2 5 and respiratory symptoms and medication use found a consistent increase in
asthmatic children (effect estimates ranging from -1.0-1.3) with less consistent evidence for an
association in asthmatic adults in cities with mean 24-h average PM2 5 concentrations ranging from 6.1 to
19.2 (ig/m3 (see Section 6.3). An evaluation of epidemiologic studies that examined specific physiologic
alterations in the respiratory health of asthmatic children (i.e., pulmonary function and pulmonary
December 2008
2-16
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
inflammation) found: (1) a decrease in forced expiratory volume (FEV,) ranging from 1-3.4% per
10 |ig/m3 increase in PM25. and (2) an increase in eNO ranging from 0.46 to 6.99 ppb, respectively. In
addition, epidemiologic studies that examined the effect of short-term exposure to PM2 5 on respiratory
HAs and ED visits found consistent associations (ranging from ~0 to 5%) for respiratory diseases (e.g.
COPD and respiratory infections) among older adults (Figure 6-20), but less consistent effects were
reported for asthma HAs and ED visits. These respiratory HA and ED visit studies were conducted in
cities with mean 24-h average PM2 5 concentrations ranging from 13.8 to 18.9 (ig/m3.
Human clinical studies provide supporting evidence of an association between short-term exposure
to PM2 5 and respiratory morbidity through increased markers of pulmonary inflammation following DE
and other traffic-related exposures, oxidative responses to DE and woodsmoke, and exacerbation of
allergic responses and allergic sensitization in response to diesel exhaust particles (DEP) (see
Section 6.3).
Toxicological studies have also demonstrated respiratory-related effects following acute PM2 5
exposure, including altered pulmonary function, mild pulmonary inflammation and injury, oxidative
responses, airway hyperresponsiveness (AHR) in allergic and non-allergic animals, exacerbations of
allergic responses and increased susceptibility to infections in a large number of studies involving
exposure to CAPs, DE, other traffic-related PM, and woodsmoke (see Section 6.3). The numerous and
wide range of respiratory responses observed in both the human clinical and toxicological studies provide
biological plausibility for an association between short-term exposure to PM2 5 and respiratory morbidity.
The consistent and coherent results found in the epidemiologic, human clinical, and toxicological
literature provide sufficient evidence to conclude that a Causal relationship is likely to exist between
short-term exposures to ambient concentrations of PM2.5 and respiratory morbidity
Mortality
An evaluation of the epidemiologic literature indicates consistent positive associations between
short-term exposure to PM2 5 and all-cause, respiratory- and cardiovascular-related mortality. The analysis
of multicity studies found that risk estimates for all-cause (non-accidental) mortality ranged from
0.29-1.21% per 10 (ig/m3 increase in PM2 5 at mean 24-h average concentrations ranging from
6.7-34.4 (ig/m3 (see Section 6.5.2.2). Cardiovascular-related mortality risk estimates (0.34-0.94%) were
found to be similar to those for all-cause mortality. However, the risk estimates for respiratory-related
mortality were slightly larger (1.01-2.2%) using the same lag and averaging indices. A regional and
seasonal pattern in PM2 5 risk estimates was observed with the greatest effects occurring in the Eastern
U.S. and during the spring. An evaluation of potential confounding of risk estimates by gaseous pollutants
found that PM2 5 mortality risk estimates remained robust to the inclusion of copollutants in regression
December 2008
2-17
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
models; however, one Canadian-based study did observe potential confounding by N02 An examination
of effect modifiers (e.g., demographic and socioeconomic status [SES] factors), specifically air
conditioning use, which is sometimes used as a surrogate for ventilation rate, suggests that PM2 5 risk
estimates increase as the percent of the population with access to air conditioning decreases. The
epidemiologic evidence, along with the results from the examination of potential confounders and effect
modifiers of the PM2 5-mortality relationship, provide sufficient evidence to conclude that a Causal
relationship is likely to exist between short-term exposure to ambient concentrations of PM2.5 and
mortality.
2.3.2.2. Effects of Long-Term Exposure to PM2.5
Size Fraction Outcome Causality Determination
PM2.5 Cardiovascular morbidity Likely to be causal
Respiratory Morbidity
Likely to be causal
Mortality
Likely to be causal
Reproductive and Developmental
Suggestive
Cancer Inadequate
Cardiovascular Morbidity
Epidemiologic and toxicological studies have provided evidence of the adverse effects of long-term
exposure to PM2 5 on clinical and subclinical markers of cardiovascular morbidity (see Section 7.2). The
epidemiologic evidence consists of a large U.S. based cohort of post-menopausal women, which reported
an association between one-year average PM2 5 concentrations (mean =13.5 |ig/nr') and MI,
revascularization, and their combination with CHD death. However, assocations were not found in a
cohort of both men and women in Germany with mean one-year average PM2 5 concentrations of
23.3 (ig/m3. The examination of subclinical markers of atherosclerosis (i.e., coronary artery calcification
[CAC], abdominal aortic calcium [AAC], and carotid intimal-medial thickness [CIMT]) through
epidemiologic studies found consistent associations with chronic exposure to PM2 5. In addition, these
studies reported a modification of atherosclerotic effects in former or current smokers and individuals
taking anti-hyperlipidemic medications.
Several toxicological studies provide evidence for the accelerated development of atherosclerosis
(i.e., increased lipid deposition, increased plaque and lesion areas) in ApoE" " mice exposed to CAPs from
December 2008
2-18
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Tuxedo, NY for 4-6 months. Increased expression of tissue factor, an important initiator of thrombosis,
was also observed in aortic plaques. Another CAPs study conducted in southern California demonstrated
increased lesion area similar to that observed by the other research groups and the effect was attributable
to ultrafine traffic PM that contained particles in the accumulation mode (0.18 In addition, long-term
exposure to CAPs from Tuxedo, NY resulted in enhanced BP responses in an animal model of
hypertension. These experimental studies provide biological plausibility for adverse cardiovascular
outcomes observed in epidemiologic studies.
A limited amount of new literature is available that examines clinical cardiovascular disease
outcomes. Although inconsistent results were reported in these long-term exposure studies, the evidence
from epidemiologic, human clinical, and animal toxicological studies that examined the cardiovascular
outcomes associated with short-term exposure to PM2.5 (discussed in Section 6.2.), supports a role for the
development of cardiovascular morbidity in response to long-term exposure to PM2 5. Based on the
consistent and coherent evidence from epidemiologic and toxicological studies that examined the
association between long-term and short-term exposure to PM2 5 and cardiovascular morbidity, sufficient
evidence is available to conclude that a causal relationship is likely to exist between long-term
exposure to ambient concentrations of PM2.5 and cardiovascular morbidity
Respiratory Morbidity
Recent epidemiologic studies conducted in the U.S. and abroad provide consistent evidence of
associations between long-term exposure to PM2 5 and respiratory symptoms, asthma, and decrements in
lung function growth in children from cities with annual average concentrations ranging from 5.0-
15.5 (ig/m3 (see Section 7.3). Subchronic and chronic toxicological studies provide some evidence of
altered pulmonary function, mild inflammation, oxidative responses and histopathological changes
including mucus cell hyperplasia and immune suppression in response to CAPs, DE, roadway air and
woodsmoke. An examination of allergic animals demonstrated AHR in response to DE, but in some cases
adaptation to prolonged exposures was observed. In addition, pre- and postnatal exposure to ambient
levels of urban particles was found to affect mouse lung development. Impaired lung development is an
important mechanism by which PM exposure may decrease lung function growth in children.
Collectively, the results from the toxicological studies provide biological plausibility for the development
of respiratory-related health effects resulting from long-term exposure to PM2 5. Overall, the consistent
and coherent evidence from epidemiologic and toxicological studies is sufficient to conclude that a
causal relationship is likely to exist between long-term exposure to ambient concentrations of PM2.5
and respiratory morbidity
December 2008
2-19
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Mortality
The new epidemiologic evidence reports a consistent association between long-term exposure to
PM25 and an increased risk of mortality (with the majority of the effects ranging from >1 to 1.20) in cities
with annual average PM25 concentrations ranging from 10.2-29 (ig/m3 (see Section 7.6). New evidence
from the Harvard Six Cities cohort study shows a relatively large reduction in mortality risk associated
with a decrease in PM2 5 concentrations. Additional analyses of the Harvard Six Cities cohort and the
American Cancer Society (ACS) study in Los Angeles suggest that previous and current studies may have
underestimated the magnitude of the PM2 5-mortality association. Overall, the consistent evidence
reported across epidemiologic studies is sufficient to conclude that a Causal relationship is likely to
exist between long-term exposure to ambient concentrations of PM2.5 and mortality
2.3.3. PM2.5 Constituents or Sources Linked to Health Outcomes
Recently, epidemiologic, human clinical and toxicological studies have begun to evaluate the health
effects associated with ambient PM constituents and sources, as opposed to PM mass. This evaluation is
conducted using a variety of quantitative methods applied to the full set of PM constituents, rather than
selecting constituents a priori. A review of the recent literature identified key studies that examined the
association between specific PM constituents, PM sources and health effects (Section 6.6). Health studies
found some associations between various PM2 5 constituents or sources, and respiratory- and
cardiovascular-related effects and mortality. However, results varied by study location, the health outcome
assessed, PM2 5 constituents considered, and PM2 5 constituents from a specific source.
Table 2-1 provides an overview of the PM source categories, along with the study-specific PM2 5
constituent groupings or tracers that comprise the sources. Also included are the PM2 5 constituents for
which an association with various health effects was found. Overall, a consistent trend or pattern that
links particular constituents or sources with specific health outcomes was not observed, but a number of
PM2 5 constituent groupings that are commonly associated with sources such as crustal/soil, salt,
secondary sulfate/long-range transport, traffic, oil combustion and woodsmoke/vegetative burning were
linked with health effects.
Comparisons among these studies are difficult because of differences in handling the data, in
approaches used to model source contributions, and in the level of experience in applying source
apportionment techniques among research groups. In an intercomparison study, several research groups
used the same data sets (which contained the composition of ambient PM2 5 and daily mortality counts)
and their choice of source apportionment models to identify PM sources (see Section 6.5.2.6.). In these
studies, when examining the association between various PM sources and mortality risk estimates, it was
December 2008 2-20 DRAFT-DO NOT QUOTE OR CITE
-------
2
3
4
5
6
7
8
9
Table 2-1. Study-specific PM2.5 factor/source categories associated with health effects.
Source Category
Health Effects
Time
Type
of
Study1
Species
Reference
CRUSTAUSOIL
Al, Si, Fe
negative association with total
mortality
Lag 2
E
Human
Mar et al.
(2000)
Al, Si, Ca, K, Fe
ST-segment depression
Lag 3
E
Human
Lanki et al.
(2006b)
Al, Si, Ti, Fe
f uric acid
f mean cycle length
Lag 15 h
E
Human
Riediker et
al. (2004b)
Al, Si, Ca, K, Fe
I ST-segment voltage
2 days post-exposure
H
Human
Gong et al.
(2003b)
Al, Si
ST-segment change
Following exposure
T
Dog
Wellenius et
al. (2003)
Al, Si
t blood PMN %
I blood lymphocytes %
t WBC
Following exposure
T
Dog
Clarke et al.
(2000)
Al
I airway irritation (penh)
During exposure
T
Dog
Nikolov et al.
(2008)
Al, Si, Ca
I lumen/wall ratio
24 h post-exposure
T
Rat
Batalha et
al. (2002)
Al, Si, Ca, Fe
4 H
t H
t SDNN, t RMSSD
During exposure
Afternoon post-exposure
Night post-exposure
T
Mouse
Lippmann
et al.
(2005b)
SALT
Na.CI
ST-segment depression
Lag 3
E
Human
Lanki et al.
(2006a)
Na.CI
f blood lymphocyte %
Following exposure
T
Dog
Clarke et al.
(2000)
Na.CI
f lung PMN density
24 h post-exposure
T
Rat
Saldiva et al.
(2002)
SECONDARY SULFATE / LONG-RANGE TRANSPORT
December 2008 2-21 DRAFT-DO NOT QUOTE OR CITE
found that the between source category variation in risk estimates for daily mortality was significantly
larger than the between group variation in reported risks. The results of this exercise indicated that the
choice of source apportionment models has a much smaller effect on variations in risk estimates
compared to the variations in risk caused by the different source components. In addition, the most
strongly associated source types were consistent across all of the groups. This study indicates that source
apportionment methods can add useful insights into those source components that contribute to PM2 5
health effects. Additionally, as more studies are conducted that increase the number of different
geographic locations in relation to similar health effects, it is probable that linkages between PM
constituents or sources and health effects will become more apparent.
-------
Source Category
Health Effects
Time
Type
of
Study1
Species
Reference
s
f total mortality
negative association with total
mortality
LagO
Lag 5
E
Human
Maret al.
(2000)
SO42- NH+», OC
f respiratory ED visits
LagO
E
Human
Sarnat et al.
(2008)
S, K, Zn, Pb
ST-segment depression
Lag 2
E
Human
Lanki et al.
(2006a)
S042"
4 systolic BP
4 h post-exposure
H
Human
Gong et al.
(2003a)
S042" (+NO2)
4FEV1
4 FVC
Following exposure
H
Human
Gong et al.
(2005)
S
| RBC
f hemoglobin
Following exposure
T
Dog
Clarke et al.
(2000)
S, Si, OC
1 H
| SDNN, | RMSSD
Afternoon post-exposure
Night post-exposure
T
Mouse
Lippmann et al.
(2005b)
TRAFFIC
Pb, Br, Cu
f mortality
Lag 0-1
E
Human
Laden et al.
(2000)
Mn, Fe, Zn, Pb, OC, EC, CO, N02
f CV mortality
Lag 1
E
Human
Mar et al.
(2000)
NOx, EC, ultrafine count
ST-segment depression
Lag 2
E
Human
Lanki et al.
(2006a)
Gasoline (OC, NOs-, NFk)
f CVD ED visits
LagO
E
Human
Sarnat et al.
(2008)
Diesel (EC, OC, NOs-)
f CVD ED visits
LagO
E
Human
Sarnat et al.
(2008)
Speed-change factor (Cu, S, aldehydes)
f blood urea nitrogen
f mean red cell volume
t blood PMN %
J, blood lymphocytes %
f von Willebrand factor (vWF)
4 protein C
f mean cycle length
t SDNN
t PNN50
f supraventricular ectopic beats
Lag 15 h
E
Human
Riediker et al.
(2004b)
Motor vehicle/other (Br, Pb, Se, Zn,
NOs-)
4 RMSSD
Afternoon post-exposure
T
Mouse
Lippmann et al.
(2005b)
Gasoline+secondary nitrate*
cytotoxic responses (potency)
24 h post-exposure
T
Rat
Seagrave et al.
(2006)
Gasoline+diesel*
inflammatory responses (potency)
24 h post-exposure
T
Rat
Seagrave et al.
(2006)
OIL COMBUSTION
V, Ni
t BALF AM %
t blood PMN %
4 blood lymphocytes %
24 h post-exposure
Following exposure
Following exposure
T
Dog
Clarke et al.
(2000)
Ni
4 respiratory rate
During exposure
T
Dog
Nikolov et al.
(2008)
V, Ni
f lung PMN density
24 h post-exposure
T
Rat
Saldiva et al.
(2002)
V, Ni, Se
4 SDNN, 4 RMSSD
Afternoon post-exposure
T
Mouse
Lippmann et al.
(2005b)
December 2008
2-22
DRAFT-DO NOT QUOTE OR CITE
-------
Source Category
Health Effects
Time
Type
of
Study1
Species
Reference
COAL COMBUSTION
Se, S042-
f mortality
Lag 0-1
E
Human
Laden et al.
(2000)
OTHER METALS
Metal processing (SO42-, Fe, NH+», EC,
OC)
f CVD ED visits
LagO
E
Human
Sarnat et al.
(2008)
WOODSMOKE / VEGETATIVE BURNING
OC, K
f CV mortality
Lag 3
E
Human
Mar et al.
(2000)
OC, EC, K, NH+»
f CVD ED visits
LagO
E
Human
Sarnat et al.
(2008)
UNNAMED FACTORS
Zn-Cu-V
f blood fibrinogen
18 h post-exposure
H
Human
Huang et al.
(2003c)
Fe-Se-sulfate
t BALF PMN
18 h post-exposure
H
Human
Huang et al.
(2003c)
Br, Pb
t BALF PMN %
24 h post-exposure
T
Dog
Clarke et al.
(2000)
Br, Pb
f lung PMN density
24 h post-exposure
T
Rat
Saldiva et al.
(2002)
Constituents not provided.
1 E = Epigemiologic study; H = Human clinical study; T = Toxicologic study
2.3.4. Public Health Impacts
2.3.4.1. PM Concentration-Response Relationship
1 The examination of the PM concentration-response curve has primarily occurred in large multi-city
2 studies that have analyzed the association between short- and long-term exposure to PM and mortality
3 (see Sections 6.5.2.7, 7.3.8, and 8.1). These studies have used various statistical methods, but overall have
4 consistently found that a no-threshold log-linear model most adequately portrays the PM-mortality
5 concentration-response relationship. However, some heterogeneity in the shape of the concentration-
6 response curve has been observed in an analysis that compared the concentration-response relationship
7 across individual cities. Therefore, although a consensus has been reached regarding the most likely shape
8 of the PM concentration-response curve in multicity analyses, uncertainty still exists surrounding the
9 PM-mortality concentration-response relationship on a city-to-city basis.
December 2008
2-23
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
2.3.4.2. Potentially Susceptible and Vulnerable Subpopulations
During the evaluation of the PM literature, numerous studies were identified that examined
whether underlying factors increased the susceptibility or vulnerability of an individual to PM-related
health effects. In this ISA, a susceptible subpopulation is defined as those individuals that might exhibit
an adverse health effect to a pollutant at concentrations lower than those needed to elicit the same
response in the general population or those individuals that might elicit a more adverse health effect at the
same concentration. A vulnerable subpopulation is defined as those individuals that might be differentially
exposed to higher concentrations of a pollutant than the general population, regardless of the health
outcome. The examination of both susceptible and vulnerable subpopulations to PM exposure allows for
the NAAQS to provide an adequate margin of safety for both the general population and sensitive
subpopulations (see Chapter 8 for a more detailed discussion).
Susceptibility Characteristics
Epidemiologic, human clinical, and toxicological studies provide evidence for a diverse group of
characteristics that could potentially increase the susceptibility of an individual to PM-related health
effects (see Table 8.1 for a comprehensive list of characteristics that could potentially increase the
susceptibility of an individual to PM-related health effects). The susceptibility characteristics examined in
the PM literature can theoretically be divided into two categories: (1) innate (e.g., age, gender,
race/ethnicity) and (2) pre-existing disease (see Section 8.2.1). Although the strength of the evidence for
each characteristic varies, the recent literature provides a basis for understanding the increased
susceptibility of an individual to PM-related health effects.
The literature provides mixed evidence that innate characteristics lead to increased susceptibility to
health effects upon exposure to PM. The evaluation of epidemiologic and human clinical studies found
some evidence, which demonstrates an increase in cardiovascular health effects in older individuals (65+)
along with some support for an increase in mortality. In addition, the epidemiologic literature suggests an
increase in respiratory-related health effects in children. When examining gender and race/ethnicity, it
remains unclear if either modifies the association between PM and health effects. The new literature
indicates no clear pattern of effect modification when stratifying effects by gender or race/ethnicity. But,
there is some evidence, albeit from two studies conducted in southern California with 6 overlapping
cities, that individuals of Hispanic ethnicity are more susceptible to mortality upon short-term exposure to
pm25.
Recent toxicological studies have also examined the effects of exposure to PM during pregnancy
through the use of animal models. These studies suggest that exposure to particles, both immunologically
December 2008
2-24
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
inert and toxic, result in local and systemic inflammation in the pregnant animal. Inflammation during
pregnancy can potentially lead to allergic susceptibility in the offspring.
The new literature has continued to examine the role of genetic factors on PM-related health
effects. These studies found some evidence that individuals with polymorphisms in genes that mediate an
antioxidant response confer a greater degree of susceptibility to PM exposure. However, it has also been
observed that in some cases genetic polymorphisms can result in a gain of function, leading to a
protective effect.
A large amount of literature examined the role of underlying diseases on PM-related health effects.
Epidemiologic, human clinical and toxicological studies have found some evidence of an increase in
cardiovascular effects in individuals with a pre-existing cardiovascular disease. However, the
cardiovascular health effect observed upon exposure to PM was found to vary depending on the pre-
existing cardiovascular condition.
The evaluation of studies that examined the association between exposure to PM and health effects
in individuals with pre-existing respiratory diseases also observed a difference in effects depending on the
underlying respiratory condition. The epidemiology literature indicated some evidence of an increase in
respiratory-related health effects in individuals with asthma, but less consistent evidence in those with
COPD. In addition, an increase in mortality was observed in individuals with underlying pneumonia and
respiratory illnesses. The toxicological literature presented evidence that suggested that individuals with
allergic airways disease are more susceptible to allergic responses upon exposure to PM. Interestingly, the
human clinical and animal toxicological literature also found evidence of cardiovascular effects (e.g.,
acute responses in the cardiovascular system and reduced pulmonary artery lumen-to-wall ratio) in
individuals or subjects with underlying respiratory illnesses.
As obesity and diabetes have increased in the U.S., studies have also examined the potential
susceptibility of these individuals to PM-related health effects. The epidemiologic literature provides
some evidence that individuals with diabetes are at increased risk of mortality along with cardiovascular
HAs and ED visits following short-term exposure to ambient concentrations of PM. In addition, human
clinical studies provide evidence of an increase in biomarkers associated with inflammation, oxidative
stress and acute phase response potentially leading to cardiovascular effects. Only a few studies have
examined the role of obesity on PM-related health effects, but there is evidence of HRV modification and
increased inflammatory markers with PM exposure.
Vulnerability
The recent epidemiologic literature has also examined characteristics that potentially increase the
vulnerability of subpopulations to PM-related health effects. These analyses, which primarily focus on the
December 2008
2-25
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
association between PM and mortality, examine health effects in individuals that are potentially
disproportionately exposed to PM due to the location of their residence, their socioeconomic status (SES),
their educational attainment, or the geographic area of the country where they live (see Section 8.2).
There is some evidence that individuals residing in an urban environment are at increased risk of
PM-related health effects, specifically mortality, due to higher exposures to traffic-derived PM. However,
when examining other factors that potentially contribute to whether an individual lives in an urban
environment, such as SES and educational attainment, it remains unclear whether either characteristic
results in disproportionate exposure to PM. This is because, specifically for SES, which is closely tied to
educational attainment, there has not been a consistently-observed trend that characterizes the impact of
SES on exposure to PM or other air pollutants. Additional analyses of air conditioning (AC) use, which is
sometimes used as a surrogate for SES, provide some evidence that AC use reduces exposure to PM.
However, it has been argued that AC use may not be an appropriate measure when examining PM
exposure because of differences in building ventilation rate, which differs by season and community.
Finally, analyses conducted that examine the PM-mortality association (for both PMi0 and PM2 5) by
geographic location tend to see the greatest effects in the Eastern U.S. Additional regional analyses have
also suggested that regional effects may vary by season with an increase in PMi0 mortality risk estimates
during the summer in the Northeast and industrial Midwest.
December 2008
2-26
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Chapter 3. Source to Human Exposure
3.1. Introduction
This chapter contains basic information about concepts and findings in atmospheric sciences and
human exposure assessment relevant to recent PM studies and results to help establish a foundation for
the detailed discussions of health effects data in subsequent chapters. Section 3.2 presents an overview of
basic information related to the size distribution and composition of airborne particles. Section 3.3
provides a brief description of the sources and emissions of PM. It includes discussions of mechanisms of
secondary PM formation from gaseous precursors. Issues related to the measurement of PM and its
components and to the deployment of monitors in networks are covered in Section 3.4. Analyses of data
for ambient concentrations of PM and its components are characterized in Section 3.5. This Section also
includes results from receptor modeling studies of source contributions to PM based on ambient data.
Policy relevant background concentrations of PM, i.e., those concentrations defined to result from
uncontrollable sources, are presented in Section 3.6. Issues related to personal exposure to PM and its
components are discussed in Section 3.7. See the 2004 PM AQCD (U.S. EPA, 2004) for a detailed
characterization of PM properties.
The intent of this chapter is to build on previous AQCDs with newly available data and studies.
This information includes new knowledge of PM chemistry, latest developments in monitoring
methodologies, recent national and local trends in PM concentration as a function of size range and
species, revised estimates of policy-relevant background PM, and recent work on exposure assessment.
This information is compiled to support interpretation of the epidemiologic studies presented in
subsequent chapters.
Developments in our ability to identify organic components in PM and in the chemistry of
formation of secondary organic aerosols indicate that oligomers are likely a major component of OC in
aerosol samples. Until a few years ago, the oxidation of terpenes and aromatic compounds were
considered as sources of SOA, but not the oxidation of isoprene. However, recent observations suggest
that small but important quantities of SOA are formed from isoprene oxidation. Gasoline engines have
been found to emit a mix of OC, EC, and nucleation mode heavy and large polycyclic aromatic
hydrocarbons (PAHs), on which unspent fuel and trace metals condense, while diesel particles are
composed of a soot nucleus on which S042 and hydrocarbons condense. After initial emission from the
vehicle, evolution of the PM distribution within the plume is a function of turbulence diluting the plume
and cycling of semi-volatile components between the gas and aerosol phases. Emissions of ultrafine
December 2008
3-1
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
particles tend to be higher during cold engine starts and when ambient temperatures are low, such as
during the winter. Current inventories of emissions from combustion sources overestimate the primary
component of organic aerosol and underestimate the semi-volatile components in the emissions. This
situation results from the lack of capture of evaporated semi-volatile components upon dilution in
standard emissions tests. Near traffic, sources of organic aerosol are underestimated. However, further
downwind the overall formation rate of SOA increases as a result of the oxidation of these semi-volatile
components. Evaluation of federal reference monitors (FRMs) and equivalent monitors strengthens
conclusions about their performance. Issues still remain regarding the recently promulgated FRM for
PMio.2.5. These data are still subject to large errors, resulting in negative numbers in a number of
locations.
In general, levels and spatial distributions of PM2.5 (based on 2005 to 2007 data) have remained
relatively unchanged since the last review (based on 1999 to 2001 data). Improvements in QAhave
resulted in slightly greater homogeneity across urban areas for PM25. Spatial variability in PM25 across
urban areas is related to differences in topography and source characteristics. Data for ultrafine particles
are still sparse, but recent studies suggest a high degree of spatial variability with motor vehicle exhaust
as the major primary source. Data from the CSN indicate West to East gradients in a number of
components. OC and EC are higher in the West than in the East, conversely sulfate is higher in the East.
Nitrate shows highest values in the valleys of central California, but similarly high values are also found
in the Midwest during winter. Studies of intra-urban variation in PM concentration have shown
neighborhood scale variability, particularly with respect to PM10 and ultrafine PM. Within a street canyon,
changes in wind direction and speed can cause significant variability over a small distance in
concentrations of PM components.
Receptor modeling studies indicate that the main sources for most of the PM2.5 found in the East
are power plants and motor vehicles. An intercomparison of source apportionment techniques indicated
that the same major source categories of PM25 were consistently identified by several independent
groups. Soil-, sulfate-, residual oil-, and salt-associated mass were most clearly identified by the groups.
Other sources such as vegetative burning and traffic related emissions were less consistently identified in
large part because of collinearity in the source profiles or the level of experience of the investigators.
Significant associations were reported between PM and certain source categories and mortality. The
between-source variation in predicting daily RR was found to be statistically significant and significantly
larger than the between group variation in this intercomparison.
Policy relevant background concentrations of PM2 5 were estimated using CMAQ nested inside a
global chemistry-transport model. On average they are <1 |_ig/m3 but daily maximum values ranged from
~3 to -63 (ig/m3 at nine National Park sites across the U.S.
December 2008
3-2
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Of major concern is the ability of PM25 and PM10 measured by ambient monitors to serve as a
reliable indicator of personal exposure to PM2 5 and PM10 of ambient origin. The key question is what
errors are associated with using PM measured by ambient monitors as a surrogate for personal exposure.
Exposure to PM of ambient origin is subject to a number of factors, including: season; region; intra-urban
spatiotemporal variability; proximity to sources; and time spent indoors and outdoors. Studies of on-street
PM exposure suggest that personal outdoor local environmental exposures to fine and coarse PM are
higher than ambient concentrations measured at urban background ambient monitors as a result of local
sources and trapping within street canyons. As a result, data reported by ambient monitors located at
background, central urban, roadside, or near-residential sites will likely be variable across an urban area.
Several recent studies have shown how the ambient component of personal exposure to PM2 5 could be
estimated using various tracer and source apportionment techniques and that it is highly correlated with
ambient concentrations of PM2 5. These studies also show that the non-ambient component of personal
exposure to PM2 5 is basically uncorrelated with ambient PM2 5 concentrations. A number of studies have
examined whether gaseous copollutants act as confounders or surrogates for PM in exposure assessments.
Many of these studies have shown that ambient 03, N02, and S02 are associated with personal exposure
to total PM2 5 and to PM2 5 of ambient origin as opposed to personal exposures to the gases, themselves.
This result may have arisen in part because personal exposures to the gases were often beneath detection
limits of the personal monitoring devices. The association between ambient 03 and ambient PM2 5 was
also generally found to be seasonal with a positive association in the summer and a negative association
in the winter. This situation arises because of seasonal differences in sources of PM2 5 and 03. The
evidence is mixed regarding the association between SES and ambient personal exposures to PM.
3.2. Overview of Basic Aerosol Properties
Unlike gas-phase pollutants such as S02, CO, H2CO and 03, which are well-defined chemical
entities, atmospheric PM varies in size, shape, and chemical composition. Atmospheric chemical and
microphysical processing of direct emissions of PM and precursors and mechanical generation of
particles tend to produce distinct lognormal modes (Whitby, 1978) as shown in Figure 3-1. To the extent
that information is available, discussions in this and subsequent chapters will focus on particles in specific
size ranges (i.e., PM2 5, PM10_2 5 and PM10). The subscripts for PM2 5 through PM10 refer to the
aerodynamic diameters in micrometers ((.im) of 50% cut points of sampling devices. For example, EPA
defines PMi0 as particles collected by a sampler with an upper 50% cut point of 10 |_im aerodynamic
diameter and a specific, fairly sharp, penetration curve, as defined in the Code of Federal Regulations (40
December 2008
3-3
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
CFR Part 58). PM25 is defined in an analogous way. Ultrafine particles, 0.01 to 0.1 |_un. thermodynamic
diameter, which is related to the diffusion coefficient of the particle, will also be discussed.
The terms "fine particles" and "coarse particles" have lost the precise meaning given in Whitby's
(1978) definition, where "fine particles" refers to all particles in the nucleation, Aitken, and accumulation
modes; and "coarse particles" characterizes all particles larger than these. Ultrafine particles correspond
loosely to the nucleation plus Aitken modes (in earlier literature, these modes were not separated and the
combination, unresolved by older instruments, was called the Aitken mode). Now, the term "fine
particles" has come to be often associated with the PM2 5 fraction, which includes the nucleation, Aitken
and accumulation modes and some of the lower-size tail of the coarse particle mode between about 1 and
2.5 |_im aerodynamic diameter. "Thoracic coarse" is frequently used to refer to PMi0.2.5, which does not
include all of the smaller coarse particles. Under high relative humidity conditions, the larger particles in
the accumulation mode may also extend into the 1 to 3 |_im size range. These relationships can be seen in
Figure 3-1, which shows the number distribution for ultrafine particles and the volume distribution (or
mass distribution if particle density is constant across the size range) for fine and (thoracic) coarse
particles. The figure is arranged this way because particle number is most highly concentrated in the
ultrafine size range but volume (or mass) is most concentrated in the larger size ranges.
Particles can penetrate to different regions of the human respiratory tract depending on size.
Thoracic particles refer to particles that travel past the larynx to reach the lung airways and the gas-
exchange region of the lung, and respirable particles are those that reach the gas-exchange region. The
selection of PM10 as an indicator of thoracic particles was based in large part on dosimetry (U.S. EPA,
1984). However, dosimetric considerations do not provide insight into the selection of a size cut to
characterize fine mode particles. The American Conference of Governmental Industrial Hygienists
(ACGIH, 2005), the International Standards Organization (ISO), and the European Standardization
Committee (CEN) have adopted a 50% cut point of 4 |_im as an indicator of respirable particles. PMi0 is
an indicator of thoracic particles; PM2 5 is an indicator of fine mode particles; and PMi0.2 5 is an indicator
of the thoracic component of coarse mode particles that is sometimes refered to as thoracic coarse (noting
that it excludes some coarse particles below 2.5 |_im and above 10
As can be seen from Table 3-1, particles in individual size modes are characterized by rather
distinct sources, composition, chemical properties, lifetimes in the atmosphere (t) and distances over
which they can travel. Whereas particles in the smaller size modes are formed mainly by combustion
processes and by nucleation and condensation of gases, coarse mode particles are generated mainly by
mechanical activity, such as by the action of wind on either the ground or the sea surface. Further details
are given in subsequent sections of this chapter.
December 2008
3-4
DRAFT-DO NOT QUOTE OR CITE
-------
Fine Particles
Coarse Particles
Ultrafine Particles
Nucleation
Mode
c? 30 —
Aitken
Mode
Droplet
Submode
Accumulation
Mode
30—
Coarse
Mode
20—
Condensation
Submode
0.1 1 10
Diameter (micrometers)
Figure 3-1. Particle size distributions by number and volume.
Source: Pandis (2004)
December 2008
3-5
DRAFT-DO NOT QUOTE OR CITE
-------
Table 3-1. Characteristics of ambient fine (ultrafine plus accumulation-mode) and coarse
particles.
Fine
Ultrafine
Accumulation
Formation
Processes
Combustion, high-temperature processes, and atmospheric reactions
Break-up of large solids/droplets
Formed by
Nucleation of atmospheric gases
including H2SO4, NH3 and some
organic compounds
Condensation of gases
Condensation of gases
Coagulation of smaller particles
Reactions of gases in or on particles
Evaporation of fog and cloud droplets in which
gases have dissolved and reacted
Mechanical disruption (crushing, grinding,
abrasion of surfaces)
Evaporation of sprays
Suspension of dusts
Reactions of gases in or on particles
Composed of
Sulfate
EC
Metal compounds
Organic compounds with very low
saturation vapor pressure at ambient
temperature
Sulfate, nitrate, ammonium, and hydrogen ions
EC
Large variety of organic compounds
Metals: compounds of Pb, Cd, V, Ni, Cu, Zn, Mn,
Fe, etc.
Particle-bound water
Bacteria, viruses
Suspended soil or street dust
Fly ash from uncontrolled combustion of
coal, oil, and wood
Nitrates/chlorides/sulfates from
HNO3/HCI/SO2 reactions with
coarse particles
Oxides of crustal elements (Si, Al, Ti, Fe)
CaC03, CaS04, NaCI, sea salt
Bacteria, pollen, mold, fungal spores, plant
and animal debris
Tire, brake pad, and road wear debris
Solubility
Not well characterized
Largely soluble, hygroscopic, and deliquescent
Largely insoluble and nonhygroscopic
Sources
High temperature combustion
Atmospheric reactions of primary,
gaseous compounds.
Combustion of fossil and biomass fuels, and high
temperature industrial processes, smelters,
refineries, steel mills etc.
Atmospheric oxidation of NO2, SO2, and organic
compounds, including biogenic organic species
(e.g., terpenes)
Resuspension of particles deposited onto
roads
Suspension from disturbed soil (e.g.,
farming, mining, unpaved roads)
Construction and demolition
Uncontrolled coal and oil combustion
Ocean spray
Atmospheric
half-life
Minutes to hours
Days to weeks
Minutes to hours
Removal
Processes
Grows into accumulation mode
Diffuses to raindrops
Forms cloud droplets and rains out
Dry deposition
Dry deposition by fallout
Scavenging by falling rain drops
Travel distance
<1 to 10s of km
100s to 1000s of km
<1 to 10s of km (small size tail, 100s to
1000s in dust storms)
Source: Adapted from Wilson and Suh (1997).
3.3. Sources of Primary and Secondary PM
1 Table 3-2 summarizes anthropogenic and natural sources for the major primary and secondary
2 aerosol constituents of fine and coarse particles. Anthropogenic sources can be further divided into
3 stationary and mobile sources. Stationary sources include fuel combustion for electrical utilities,
4 residential space heating and cooking; industrial processes; construction and demolition; metals, minerals,
5 and petrochemicals; wood products processing; mills and elevators used in agriculture; erosion from tilled
6 lands; and waste disposal and recycling. Mobile or transportation-related sources include direct emissions
December 2008
3-6
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
of primary PM and secondary PM precursors from highway vehicles and non-road sources as well as
fugitive dust from paved and unpaved roads. In addition to fossil fuel combustion, biomass in the form of
wood is burned for fuel. Vegetation is burned to clear new land for agriculture and for building
construction, to dispose of agricultural and domestic waste, to control the growth of animal or plant pests,
and to manage forest resources (prescribed burning). Wildlands also burn due to lightning strikes and
arson. Also shown in Table 3-2 are sources for several precursor gases, the oxidation of which can form
secondary PM.
In general, the sources of fine PM are very different from those of coarse PM. Some of the mass in
the fine size fraction forms during combustion from material that has volatilized in combustion chambers
and then recondensed before emission to the atmosphere. Some ambient PM2.5 forms in the atmosphere
from photochemical reactions involving precursor gases. PM formed by the first mechanism is referred to
as primary, and PM formed by the second mechanism is referred to as secondary. PMi0.2.5 is mainly
primary in origin, as it is produced by surface abrasion or by suspension of biological material composed
of microorganisms (e.g., bacteria, viruses, fungal spores, pollens) and fragments of living things
(e.g., plant and insect debris). Because precursor gases undergo mixing during transport from their
sources and reactions in the atmosphere can produce the same products, it is difficult to identify
individual sources of secondary PM constituents. Transport and transformation of precursors can occur
over distances of hundreds of kilometers. Coarse PM has a shorter lifetime in the atmosphere, so its
effects tend to be more localized.
Only major sources for each constituent within each broad category shown at the top of Table 3-2
are listed. Not all sources are equal in magnitude. Chemical characterizations of primary particulate
emissions for a wide variety of natural and anthropogenic sources (as shown in Table 3-2) were given in
Chapter 5 of the 1996 PM AQCD (U.S. EPA, 1996). Summary tables of the composition of source
emissions presented in the 1996 PM AQCD (U.S. EPA, 1996) and updates to that information are
provided in Appendix 3D to the 2004 PM AQCD (U.S. EPA, 2004). Source composition profiles are
archived by the EPA at http://www.epa.gov/ttn/chief/software/speciate/index.html. The profiles of source
composition were based in large measure on the results of studies that collected source signatures for use
in source apportionment studies.
Natural sources of primary PM include windblown dust from undisturbed land, sea spray, and
biological material. The oxidation of a fraction of terpenes emitted by vegetation and reduced sulfur
species from anaerobic environments leads to secondary PM formation. Ammonium (NH4) ions, which
play a major role in regulating the pH of particles, are derived from emissions of NH3 gas. Source
categories for NH3 have been divided into emissions from undisturbed soils (natural) and emissions that
are related to human activities (e.g., fertilized lands, domestic and farm animal waste). There is ongoing
December 2008
3-7
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
debate about characterizing emissions from wildfires (i.e., unwanted fire) as either natural or
anthropogenic. Wildfires have been listed in Table 3-2 as natural in origin, but land management practices
and other human actions affect the occurrence and scope of wildfires. For example, fire suppression
practices allow the buildup of fire fuels and increase the susceptibility of forests to more severe and
infrequent fires from whatever cause, including lightning strikes. Similarly, prescribed burning is listed as
anthropogenic, but can be viewed as a substitute for wildfires that would otherwise occur eventually on
the same land.
Information about the nature of sources directly emitting ultrafine particles is still sparse compared
to that for the larger size modes. The composition of motor vehicle has been most widely studied. Matti
Maricq (2007) presents a model of diesel PM as a mix of nucleation-mode S042 and hydrocarbons from
unspent fuel and soot embedded with trace metals on which S042 and hydrocarbons condense, although
the S042 portion of the diesel PM is now substantially reduced following mandate of ultra-low sulfur
diesel fuel production (Lim et al., 2007b). Similarly, gasoline PM is a mix of OC, EC and small quantities
of trace metals and sulfates, with OC constituting anywhere from 26-88% of PM (Cadle et al., 1999)
(Geller et al., 2006; Schauer et al., 2002). Much of the OC mass has yet to be identified. Riddle et al.
(2007) have shown that large and heavy PAHs are present in OC. Phuleria et al. (2006) have also found
spikes of high molecular weight and large PAHs from gasoline-fueled vehicles in the Caldecott Tunnel.
The diameter of semi-volatile near-road PM is highly dependent on the temperature at which the particles
exit the engine and the ambient temperature, so that hot driving conditions and/or hot ambient conditions
can cause the mode of the ultrafine size distribution to decrease in diameter and magnitude (Kuhn et al.,
2005b).
An overview of estimates of emissions of primary PM and precursors to secondary PM from major
sources is given in Section 3.3.1. The transformations from gaseous precursors to secondary PM shown in
Table 3-2 are described in Section 3.3.2.
December 2008
3-8
DRAFT-DO NOT QUOTE OR CITE
-------
Table 3-2. Constituents of atmospheric particles and their major sources.
Primary (PM <2.5 |jm)
Primary (PM >2.5 |jm)
Secondary PM Precursors
(PM <2.5 jjm)
Aerosol species
Natural Anthropogenic
Natural
Anthropogenic
Natural
Anthropogenic
Sulfate (S042-)
Sea spray
Fossil fuel
combustion
Sea spray
Oxidation of
reduced sulfur
gases emitted by
the oceans and
wetlands and SO2
and H2S emitted
by volcanism and
forest fires
Oxidation of SO2
emitted from fossil
fuel combustion
Nitrate (NO3")
Oxidation of NOx
produced by soils,
forest fires, and
lighting
Oxidation of NOx
emitted from fossil
fuel combustion and
in motor vehicle
exhaust
Minerals
Erosion and re-
entrainment
Fugitive dust from
paved and
unpaved roads,
agriculture,
forestry,
construction, and
demolition
Erosion and re-
entrainment
Fugitive dust,
paved and
unpaved road
dust, agriculture,
forestry,
construction, and
demolition
Ammonium (NPU*)
Emissions of NH3
from wild animals,
and undisturbed
soil
Emissions of NH3
from animal
husbandry, sewage,
and fertilized land
Organic carbon (OC)
Wildfires
Prescribed
burning, wood
burning, motor
vehicle exhaust,
cooking and
industrial
processes
Soil humic matter
Tire and asphalt
wear, paved and
unpaved road
dust
Oxidation of
hydrocarbons
emitted by
vegetation
(terpenes, waxes)
and wild fires
Oxidation of
hydrocarbons
emitted by motor
vehicles, prescribed
burning, wood
burning, solvent use
and industrial
processes
EC
Wildfires
Motor vehicle
exhaust (mainly
diesel), wood
biomass burning,
and cooking
Tire and asphalt
wear, paved and
unpaved road
dust
Metals
Volcanic activity
Fossil fuel
combustion,
smelting and other
metallurgical
processes,and
brake wear
Erosion, re-
entrainment, and
organic debris
Bioaerosols
Viruses and
bacteria
Plant and insect
fragments, pollen,
fungal spores, and
bacterial
agglomerates
Dash (—) indicates either very minor source or no known source of component.
Source: U.S. EPA (2004).
3.3.1. Emissions of Primary PM and Precursors to Secondary PM
Emissions of primary PM2 5, PMi0 and gaseous precursor species (S02, NOx, NH3 and VOCs) from
different source categories are shown in Figure 3-2. Note that the entries refer mainly to anthropogenic
December 2008
3-9
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
sources, with little information about natural sources. However, for categories such as VOCs, the
contribution from biogenic emissions of isoprene and terpenes can be quite large. The entries are
continually undergoing revision and are subject to varying degrees of uncertainty. For example, almost all
of the sulfur in fuel is released as volatile components (S02 or S03) during combustion. Hence, sulfur
emissions can be calculated on the basis of sulfur content in fuel stocks to greater accuracy than can be
done for other pollutants like nitrogen oxides or primary PM. There have been notable downward
revisions to the inventories since 2002 in the emissions of dust from roads. These have resulted in large
measure from incorporation of emissions test data that relies on updated measurement methods. Also, the
spatial and temporal nature of wildland fire emissions has improved since 2002 by integrating satellite-
derived fire detection and state-of-art fuels characterization and consumption models (Pouliot et al.,
2008). Many estimates for emissions from high temperature combustion sources are subject to artifacts
due to inadequate dilution and thermal equilibration of the samples (England et al., 2007a; England et al.,
2007b; Sheya et al., 2008). Note that the estimates given in Figure 3-2 are U.S. national averages and may
not accurately reflect the contribution of specific local sources determining a person's exposures to PM at
any given time and location.
December 2008
3-10
DRAFT-DO NOT QUOTE OR CITE
-------
Fertilizer &
bo ¦- ent
Livestock
Road Dust
WHlfii
c :cl
!ncust
On-Road
2%
nrtvP.es
Non-Road ,
Industrie
Waste fertilizer &
Disposal-
Livestock
PM? s (5.4 MMT)
Residential
On-Road Wood ,
Non-Road 2% \ q%
Road LAJSt
Sc vent
. 3 c> t
DisposaE
Industna
Must
CCTTY.-teS
Fertilizer &
Livestock
Nsn-Road On-Road
0% "\ 7%
Msc.
1%
Industrial
Rocesses -
5%
hdustf
Corrm'Res
Fuels
1% j
Fires /
6%
S02 (13.9 MMT)
Residential
Wood
0%
Solvent
- Use
Road CXist ... qo/o
0% Waste
Disposal
fertilizer &
Livestock
Solvent
8% r
Road Dust
49%
NH3 (3.8 MMT)
Comm'Res
Fuels
4%
industrial
Recesses
5%
Msc.
25%
Non-Road
Residential
Wood
2%
2%
On-Road
1%
PM10 (19,9 MMT)
Solvent
Waste
Disposal
1%
EG Us
22% Fertilizer &
i- Livestock
I 0%
Head _• st
Residential
'.Yccc
On-Road
hdust'
Non-Road
Industrial
Rocesses
Msc. S%
0%
NO* (19,4 MMT)
Ferti izer &
¦¦ .lists
Livestock
So'vent
Fires industf
19% Comm'Res
Fuels
1%
Roaa LXist
housing
Residential
Wood
On-Road
Non»nMd
VOC (18.6 MMT)
Source: U.S. EPA (2006a)
Figure 3-2. Detailed source categorization of emissions of primary PM2.5, PM10 and gaseous
precursor species SO2, NOx, NH3 and VOCs for 2002 in units of million metric tons.
E<3TJs «= Electricity Generating I "nits.
December 2008
3-11
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
3.3.2. Formation of Secondary PM
Precursors to secondary PM have natural and anthropogenic sources, just as primary PM has
natural and anthropogenic sources. A substantial fraction of the fine particle mass, especially during the
warmer months of the year, is secondary in nature, formed as the result of atmospheric reactions
involving both inorganic and organic gaseous precursors. The major atmospheric chemical
transformations leading to the formation of particulate nitrate (pN03) and sulfate (pS04) are relatively
well understood; whereas those involving the formation of SOA are less so and are subject to much
current investigation. A large number of organic precursors are involved and many of the kinetic details
still need to be determined. Also, many of the actual products of the oxidation of hydrocarbons have yet to
be identified. However, there has been substantial progress made in understanding the chemistry of SOA
formation in the past few years.
3.3.2.1. Formation of Nitrate and Sulfate
The basic mechanism of the gas and aqueous phase oxidation of N02 and S02 has long been
studied and can be found in numerous texts on atmospheric chemistry, e.g., Seinfeld and Pandis (1998),
Finlayson-Pitts and Pitts (2000), Jacob (1999), and Jacobson (2002a). The reader is referred to the 2004
PM AQCD (U.S. EPA, 2004), where these processes are described in great detail, as well as the 2008
NOx ISA (U.S. EPA, 2008c) and the 2008 SOx ISA (U.S. EPA, 2008d).
3.3.2.2. Formation of Secondary Organic Aerosol
Some key new findings have altered perceptions of secondary organic aerosol (SOA) formation
since the 2004 PM AQCD (see especially the reviews by Kroll and Seinfeld (2008) and Rudich et al.
(2007). Robinson et al. (2007a) noted that emissions from combustion sources overestimates the primary
component of organic aerosol and underestimates the semi-volatiles in the emissions. This situation
results from the lack of capture of evaporated semi-volatile components upon dilution in standard
emissions tests. Near traffic, sources of organic aerosol are underestimated. However, further downwind
the overall formation rate of SOA increases as a result of the oxidation of these semi-volatile components.
Until a few years ago, the oxidation of terpenes and aromatic compounds were considered as sources of
SOA, and the oxidation of isoprene was not considered as a source of SOA. However, observations of
2-methyl tetrols in ambient samples from a number of different environments suggest that small but
important quantities of SOA are formed (Claeys et al., 2004). Laboratory studies also indicate the
formation of 2-methyl tetrols from isoprene oxidation (Edney et al., 2005; Kleindienst et al., 2006).
December 2008
3-12
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Kroll and Seinfeld (2008) and Rudich et al. (2007) stressed that the composition of SOA evolves
from repeated cycles of volatilization and condensation of chemical reaction products in both the particle
and gas phases. Rudich et al. (2007) focused on the oxidation of particle phase species by reaction with
gas phase oxidants. Kroll and Seinfeld (2008) identified three factors that determine the SOA forming
potential of organic compounds in the atmosphere:
1. Oxidation reactions of gas-phase organic species. These species include alkanes, alkenes,
aromatics, cyclic olefins, isoprene and terpenes. Note that oxidation reactions can either
lower volatility by addition of functional groups or increase volatility by cleavage of
carbon-carbon bonds;
2. Reactions in the particle, or condensed, phase that can change volatility either by oxidation or
formation of high-molecular-weight species. These reactions can lead to the formation of
oligomers e.g. to decrease volatility or to the formation of products that can volatilize; and
3. Ongoing reactions that result from the varied volatility of oxidation products.
Other detailed work has focused on the formation of higher molecular weight particle-phase
oligomers (Gao et al., 2004b; Kalberer et al., 2004; Tolocka et al., 2004), the importance of cloud
processing in the evolution of SOA (Blando and Turpin, 2000; Gelencser and Varga, 2005), and the role
of acid seeds in oligomer formation (Tolocka et al., 2004). These results imply that ambient samples could
contain mixtures of SOA from different sources at different stages of processing, some with common
reaction products making source identification of SOA problematic. Figure 3-3 shows a schematic of
processes involved in the formation of SOA.
December 2008
3-13
DRAFT-DO NOT QUOTE OR CITE
-------
PRIMARY SECONDARY
products
products
products
Droplet
evaporation
Organic
Compounds
Aerosol
phase
reactions
Gas/particle partitioning
Nucleation, sorption, condensation/evaporation
Gas phase
reactions
of alkanes,
aromatics, alkenes,
olefins, etc
and -OH, 03, NOz
Aqueous
phase
reactions
of dicarbonyls,
organic acids
and -OH, -N02, O-
Figure 3-3. Primary emissions and formation of secondary organic aerosol through gas, cloud
and condensed phase reactions.
3.4. Monitoring Issues
3.4.1. Ambient Measurement Techniques
3.4.1.1. Federal Reference Method and Federal Equivalent Method Evaluation
1 The FRM and Federal Equivalent Method (FEM) are designed to measure the mass concentrations
2 of ambient particles. The FRMs for measuring PMi0 and PM2 5 are specified in CFR 40 Part 50, Appendix
3 J and L, respectively. The PM10 FRM is a performance based method in which particles are inertially
4 separated with a penetration efficiency of 50% at 10 ± 0.5 (j,m aerodynamic diameter. The collection
5 efficiency specified in the CFR approaches 100% as particle size decreases and approaches 0% as particle
6 size increases. Particles are collected on filters for which mass concentrations are determined
7 gravimetrically. In contrast, the FRM for PM2 5 is a design based method that specifies certain details of
8 the sampler design, as well as of sample handling and analysis, whereas other aspects (e.g., flow control)
9 have performance specifications (U.S. EPA, 2004). PMi0_2.5 concentration is computed as the difference
10 between concurrent and co-located PMi0 (as specified in 40 CFR Part 50, Appendix O) and PM2 5
December 2008
3-14
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
concentrations obtained from select co-located FRM samplers. FRMs and FEMs for PM2 5, PM10, and
PM10_2 5 were discussed in detail in the 2004 PM AQCD (U.S. EPA, 2004). Issues discussed there include
the definition or description of FRMs and FEMs for PM10, PM25 and PM10.2.5 with detailed descriptions of
the WINS impactor, virtual and cascade multi-stage impactors for PMi0.2.5 measurement, and real-time or
continuous FEMs for PMi0 and PM2 5 including:
¦ Tapered Element Oscillating Microbalance (TEOM) operated at various temperatures;
¦ Sample Equilibration System (SES)-TEOM;
¦ Differential TEOM,
¦ Beta-Gauge Techniques (BGT);
¦ Piezoelectric Microbalance;
¦ Real-Time Total Ambient Mass Sampler (RAMS);
¦ Continuous Ambient Mass Monitor (CAMM);
¦ Continuous Coarse Particle Monitor (CCPM);
¦ Micro-orifice Uniform Deposit Impactor (MOUDI);
¦ Multichannel diffusion denuder sampling system (BOSS); and
¦ Light scattering photometric instruments.
In 2006, EPA finalized new performance criteria (40 CFR Part 53) for the approval of FEMs as
Class II equivalent methods when based on integrated filter sampling and as Class III equivalent methods
when based on continuous technologies that can provide at least hourly data reporting. The performance
criteria include evaluating additive bias (intercept) and multiplicative bias (slope) as well as correlation
from co-located candidate and FRM methods at field studies covering multiple seasons and locations. As
a result of these new performance criteria, EPA approved the first Class III FEM for PM2 5 on March 12,
2008. The approved method is the Met One BAM 1020 incorporating the same techniques described in
the 2004 PM Criteria Document; however, several features have been improved such as the distance of
the beta source to the filter tape and the conditioning of the sample stream relative humidity for better
sensitivity and comparison to filter-based methods. A complete list of FRMs and FEMs can be found at
the EPA web site http://www.epa.gov/ttn/amtic/files/ambient/criteria/reference-eauivalent-methods-
list.pdf and in Annex A. This ISA will focus on new methodologies and evaluation techniques.
Several new innovations have emerged since 2004 to measure both fine and coarse PM fractions in
the ambient air. These techniques include approval of a very sharp cut cyclone (VSCC) as a PM2 5 FEM
December 2008
3-15
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
method (Kenny et al., 2004), a high-volume dichotomous sampler (1000 1/min) (Sardar et al., 2006), and a
Filter Dynamics Measurement System-TEOM (FDMS-TEOM) (Grover et al., 2006). The VSCC provides
superior performance over long sampling periods under heavy loading and was also incorporated as an
optional second-stage separator for the PM 2.5 FRM (71 FR 61214, October 17, 2006). The
FDMS-TEOM system is a self-referencing system operated at 4°C integrated with the measurement
capability of the TEOM technology operated at 30°C. With oscillating measurements of 6 minutes each
for ambient air and chilled clean air - under which conditions the volatile PM loses mass on the TEOM
filter and is therefore quantified - the instrument provides measurements of total particulate mass,
including the semivolatile NH4NO3 and organic material. Several particle sizing instruments have also
been developed to measure PM concentration and particle size; these are discussed later in Section 3.4.1.3
along with ultrafine particles and PM size distribution.
Evaluation of all the instruments mentioned above was conducted both in supersite studies and in
other research studies (Ayers, 2004; Brown et al., 2006; Butler et al., 2003; Cabada et al., 2004; Chang
and Tsai, 2003; Charron et al., 2004; Chow et al., 2006; Grover et al., 2005; Hains et al., 2007; Hering et
al., 2004; Jaques et al., 2004; Krieger, 2007; 2005b; 2005c; Lee et al., 2005d; Price et al., 2003; Rees et
al., 2004; Russell et al., 2004; Salminen and Karlsson, 2003; Schwab et al., 2004; 2006; Solomon et al.,
2003; Tsai et al., 2006a; Vega et al., 2003; Wilson et al., 2006; Yi et al., 2004; Zhu et al., 2007) (Annex
A). In general, the co-located FRMs showed very good precision with CV <5%. For different co-located
FRMs, the regression slope of one sampler on another is commonly close to unity and with an R2 >0.95.
The PM2 5 and PM10 concentrations measured by dichotomous samplers were within 10% of the FRM
methods, and the differences can be attributed to the sampling artifacts of semi-volatile components; see
Section 3.4.1.2 for details. The precision of various TEOMs ranges from 10%-30%. The concentration
measured by the TEOM operated at 50°C was consistently lower than those measured by the TEOM
operated at 30°C. The TEOM operated at 30°C provided concentrations 50% lower than the
FDMS-TEOM, and the FDMS-TEOM provided concentrations 10-30% higher than the FRM. The
differences between these monitors were also found to be a function of season and location. BGTs were
highly correlated with FRMs but BGT mass could be higher than the FRM mass (30% higher at the
Fresno supersite) (Chow et al., 2008). CAMMs didn't show a consistent pattern when compared with
FRMs. These differences could be largely attributed to the sampler operating principles and design,
ambient conditions (copollutants and meteorological parameters), and the built-in default calibration
factors for non-FRM/FEM instruments. Additionally, a number of new techniques have been developed to
reduce positive and negative sampling artifacts. These are described in the ISA for NOx and SOx -
Environmental Criteria (U.S. EPA, 2008e).
December 2008
3-16
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Several papers (Buser et al., 2007a, b, c) published since the 2004 PM AQCD claim that the EPA
FRM samplers for PM10 (and PM2 5) "oversample certain agricultural and other source emissions." These
claims are based on the erroneous assumption that the "true" PM10 concentration is what would be given
by a PMio sampler that excluded all particles greater than 10 |_im aerodynamic diameter and included all
particles less than 10 |_im. As noted earlier (Section 2.2) the legal definitions for PM2 5 and PMi0, as
defined in the Code of Federal Regulations include both a 50% cut-point and a penetration curve. For
PM10, the 50% cut-point of 10 (.im diameter means that 50% of particles with aerodynamic diameter of
10 ± 0.5 |_im are removed by the inlet and 50% pass through the inlet and are collected on the filter. The
penetration curve specifies, as a function of particle size, the fraction of particles larger than 10 (.im that
pass through the inlet and the fraction of particles less than 10 |_im that are intercepted by the inlet. No
effort was made in the development of the FRM to have the PMi0 sampler collect all particles less than
10 |am and no particles greater than 10 |_im since the sampler was designed to collect a fraction of
atmospheric particles similar to the "inhalable" or thoracic fraction, i.e., those particles that would pass
through the nose and throat and reach the lungs (Miller et al., 1979). Thus, the FRM PMi0 sampler
correctly and intentionally collects particles greater than 10 (.im.
3.4.1.2. PM Speciation
The following sections describe recent developments regarding measurement techniques to
ascertain quantities of particle-bound water, cations and anions, elemental composition, carbon, and
organic species.
Particle-Bound Water
Particle-bound water is an important component of ambient PM (U.S. EPA, 2006d). Recently, a
differential method was developed to measure particle-bound water (Santarpia et al., 2004; Stanier et al.,
2004). The dry ambient aerosol size spectrometer (DAASS) can measure particle-bound water in the
particle size range from 3 nm-10 (.im (Stanier et al., 2004), by alternatively measuring ambient PM size
distribution at low relative humidity (RH) and ambient RH. A comparison of the two size distributions
provides information on the water absorption and change in particle size due to RH. Khlystov et al.
(2005) reported that the particle-bound water, measured by DAASS, was underestimated for particles
<200 nm, and overestimated for particles >200 nm, compared with thermodynamic models. The loss of
semi-volatile components during measurement may also bias the measurement results. Methods and
analytical specifications for particle-bound water are listed in Annex A.
December 2008
3-17
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Cation and Anion
The measurement of cations and anions including S042 , N03 . NH4+, CI", Na+, and K+ still heavily
relies on filter-based collection (denuders might be used in the sampling system to adjust sampling
artifacts), water based extraction and ion chromatography (IC) based chemical speciation and
quantification. These methods have been reviewed in the 2004 PM AQCD (U.S. EPA, 2004).
Filter-denuder based integrated sample methods for S042 , N03 . and NH4+ have been detailed in the 2008
SOx ISA and 2008 NOx-SOx ISA (U.S. EPA, 2008d, e).
Recent developments in multiple ion measurements have focused on the coupling of IC and a
sample dissolution system, represented by the Particle into Liquid Sampler-Ion Chromatography
(PILS-IC) and the Ambient Ion Monitor (AIM) (Dasgupta et al., 2007; Orsini et al., 2003; Weber et al.,
2001). When ambient PM passes through the PILS-IC system, water droplets are generated by mixing
ambient PM with saturated water vapor and collected by impaction. The resulting liquid stream is then
introduced into the IC system for ion speciation and quantification. Hourly concentrations of multiple
ions can be obtained with the system, with a coefficient of variation of 10%. For the AIM system, a
parallel plate denuder is used to remove the interfering gases, and then particles enter a super-saturation
chamber to form droplets. The collected droplets are then introduced into the IC for analysis. The AIM
system can provide hourly concentrations for multiple ions. The particle mass spectrometer is another
advance in multiple PM component measurements, but most of these types of measurements are
semi-quantitative and will be detailed later in Section 3.4.1.4. Note that measurement and analytical
specifications for ions other than N03 and S042 are listed in Annex A.
Sulfate
Methods used for continuous (sampling interval of minutes) measurements of S042 include
Aerosol Mass Spectrometry (AMS) (Drewnick et al., 2003; Hogrefe et al., 2004), PILS (Weber et al.,
2001), flash volatilization techniques (Bae et al., 2007a; Stolzenburg and Hering, 2000), and the Harvard
School of Public Health (HSPH) tube furnace to convert S042 to S02 for detection by a S02 analyzer
(Allen et al., 2001). These methods are described in detail by Drewnick et al. (2003), along with an inter-
sampler comparison that found overall agreement within 2.9% for all continuous instruments with R2 of
0.87 or better. When compared with filter samples, Drewnick et al. (2003) showed differences were less
than 25% for the AMS, PILS, flash vaporization, and HSPH continuous S042 monitors. Annex A lists
detailed methods and analytical specifications for sampling S042 .
December 2008
3-18
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Nitrate
In addition to the filter-based method and the new developments mentioned above, methods based
on flash volatilization-chemiluminescence analysis and catalytic conversion-chemiluminescence analysis
have also been developed for continuous N03 measurement (averaging time 30 s-10 min). For the flash
volatilization system (Fine et al., 2003; Stolzenburg and Hering, 2000; Stolzenburg et al., 2003), particles
are collected by a humidified impaction process and analyzed in place by flash vaporization and
chemiluminescent detection of the evolved NOx. For the catalytic conversion-chemiluminescence
analysis system (e.g. ARAN, Weber et al., 2003), N03 was measured by conversion of particle N03 into
NO, and then detected with the chemiluminescence method. Field and lab comparisons were conducted to
compare the different instruments mentioned above. Although the R&P 8400N ambient particulate N03
monitor could provide 10-min resolution data and showed excellent precision (with a CV <10%)
(Harrison et al., 2004; Hogrefe et al., 2004; Long and McClenny, 2006; Rattigan et al., 2006), it
consistently reported N03 concentrations -30% lower than the denuder-fllter systems in both the
Baltimore supersite and the multi-year field sampling in New York (Harrison et al., 2004; Hogrefe et al.,
2004; Rattigan et al., 2006). In the New York measurement campaign, an AMS was also co-located with
other instruments to obtain the real-time N03 information. AMS did not always agree well with the
denuder-filter system for reasons not entirely apparent. However, Bae et al. (2007b) reported that some
organic compounds can also produce signals at mass-to-charge ratio m/z = 30, which is one of the
characteristic m/z for N03 . Therefore, the disagreement between the AMS and the filter-based method
could be a result of the interference of organic compounds using the AMS. Annex A lists methods and
analytical specifications for sampling N03 .
Ammonium
Field and lab comparisons were conducted to compare the continuous measurement methods
mentioned above. Several continuous and semi-continuous instruments could be applied to monitor
ambient ammonium concentrations (Al-Horr et al., 2003; Bae et al., 2007a). Bae et al. (2007a) conducted
an inter-comparison of three semi-continuous instruments during the New York multi-year air sampling
campaign: a PILS-IC, an AMS, and a wet scrubbing-long path absorption photometer. Bae et al. (2007a)
reported the inter-sampler coefficients of determination (R2) between these instruments were above 0.75
and the slopes (intercepts were forced to be zero) were between 0.71 and 1.04. Annex A describes
measurement of ions other than N03 and S042 , including NH/.
December 2008
3-19
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Elemental Composition
Techniques for measuring the elemental composition of PM samples were reviewed in the 2004
PM AQCD (U.S. EPA, 2004). These methods include:
¦ Energy dispersive X-ray fluorescence (ED-XRF);
¦ Synchrotron X-ray fluorescence (SXRF);
¦ Particle-induced X-ray emission (PIXE);
¦ Particle elastic scattering analysis (PESA);
¦ Total reflection X-ray fluorescence (TR-XRF);
¦ Instrumental neutron activation analysis (INAA);
¦ Atomic absorption spectrophotometry (AAS);
¦ Inductively-coupled plasma-atomic emission spectroscopy (ICP-AES);
¦ Inductively-coupled plasma-mass spectrometry (ICP-MS); and
¦ Scanning electron microscopy (SEM).
Recent development in this area focused on the semi-continuous measurement methods, in which
elements were analyzed in the lab using the methods mentioned above on time-resolved and/or size
resolved samples (Kidwell and Ondov, 2004). The concentrated slurry/graphite furnace atomic absorption
spectrometry (GFAAS) method collects ambient PM as a slurry using impactors and then the collected
PM is analyzed by AAS in the lab. Laser induced breakdown spectroscopy (LIBS) was used to measure
seven metals at the Pittsburgh supersite. LIBS concentrates ambient PM using a virtual impactor into a
sample cell, and then a Nd: YAG laser-spectrometer is used to identify and quantify different elements. A
full listing of measurement techniques and analytical specifications for trace elements is provided in
Annex A.
Elemental and Organic Carbon
The large variety of aspects of carbon analyses were reviewed in the 2004 PM AQCD (U.S. EPA,
2004). Measurement and analytical specifications for carbon measurements are listed in Annex A. Aspects
of the measurements include sampling artifacts associated with the integrated filter-based OC and EC
sampling methods, the IMPROVE vs. the National Institute for Occupational Safety and Health (NIOSH)
thermal optical protocols (i.e., different thermal optical methods) and optical techniques to measure
light-absorption or BC. One significant change taking place in the CSN is that the method for carbon
measurements began switching from the NIOSH method to the IMPROVE carbon analysis protocol in
December 2008
3-20
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
2007. This is a phased process and the CSN will not be fully converted to IMPROVE-like sampling and
IMPROVE analysis until late 2009 (Henderson, 2005). The CSN network was implemented to support the
PM2 5 NAAQS and began providing data for PM2 5 mass, S042 . N03 . NH/, Na, K, EC, OC, and select
trace elements (A1 through Pb) at many sites across the U.S. This move increased consistency between
these two networks. Also, since the release of the 2004 PM AQCD, more studies have been conducted to
extend the understanding of sampling artifact issues (Chow et al., 2008; Watson et al., 2005b), to evaluate
different thermal and optical procedures (Chen et al., 2004a; Chow et al., 2004; 2005a; 2007; Conny et
al., 2003; Han et al., 2007; Subramanian et al., 2006; Watson et al., 2005b ), to develop reference
materials (Klouda et al., 2005; Lee, 2007), to develop water soluble organic carbon (WSOC)
measurement techniques (Andracchio et al., 2002; Yang et al., 2003b), and
semi-continuous/continuous/real-time carbon measurement techniques (Chow et al., 2008; Watson et al.,
2005b), as well as to introduce isotope identification into the OC/EC measurement (Huang et al., 2006).
OC sampling artifact issues were further addressed in various studies (Arhami et al., 2006; Bae et
al., 2004; Chow et al., 2005a; Fan et al., 2003; Fan et al., 2004; Grover et al., 2008; Lim et al., 2003;
Mader et al., 2003; Matsumoto et al., 2003; Miiller et al., 2004; Offenberg et al., 2007; Olson and Norris,
2005; Park et al., 2006a; Rice, 2004; Subramanian et al., 2004; ten Brink et al., 2004; 2005; Viana et al.,
2006), and were well summarized by Watson et al. (2005b) and Chow et al. (2008). There are two
commonly used methods to correct OC sampling artifacts: the filter with backup filter system (TBQ:
placing a backup quartz-fiber filter behind the front Teflon-membrane; or QBQ: placing a backup
quartz-fiber filter behind the front quartz-fiber filter), and the denuder-filter-adsorbent system.
Subramanian et al. (2004) and Chow et al. (2006) reported that during the Pittsburgh and Fresno supersite
studies the positive artifact (organic gases condense on filters) from TBQ (24-34%, up to 4 (ig/m3 OC)
was nearly twice that from QBQ (13-17%). With the denuder-filter-adsorbent system, the negative
artifact (OC evaporating from the filter) was 5-10%. Watson and Chow (2002) reported that the
XAD-coated denuder could function as efficiently as a parallel plate denuder using carbon-impregnated
charcoal filters (CIF) with frequent denuder changes. Huebert and Charlson (Huebert and Charlson)
reported that using tandem filter packs may hinder a quantitative analysis of the artifacts.
Different temperature protocols and optical correction methods in thermal-optical analyses were
further evaluated by Watson et al. (2005b), Chow et al. (2005a; 2007), Subramanian et al. (2006), Conny
et al. (2003), Han et al. (2007), Chen et al. (2004a), and Chow et al. (2004). Solomon et al. (2003)
reported a 20-50% difference for OC and a 20-200% difference for EC using 11 filter samples and 4
different analytical protocols. In an assessment of the different thermal-optical analysis protocols used
around the world, Watson et al. (2005b) reported that differences of a factor of 2 to 7 in EC between
different methods could be observed, and a factor of 2 was common, while the relative differences in OC
December 2008
3-21
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
between different methods were relatively small. As Watson et al. (2007) stated, there are 12 major
differences among the thermal methods: (1) analysis atmosphere; (2) temperature ramping rates; (3)
temperature plateaus; (4) residence time at each plateau; (5) optical pyrolysis monitoring configuration
and wavelength; (6) standardization; (7) oxidation and reduction catalysts; (8) sample aliquot and size; (9)
evolved carbon detection method; (10) carrier gas flow through or across the sample; (11) location of the
temperature monitor relative to the sample; and (12) oven flushing conditions. Chow et al. (2004) and
Chen et al. (2004a) addressed the difference between optical transmission and optical reflectance methods
for charring correction, and they reported that the charring OC on the surface of a filter or inside a filter
dominated the differences between these two charring correction methods. The differences between
different sampling and measurement methods are also applied to the in-situ/semi-continuous methods,
since most of these methods are also based on thermal-optical analysis of collected filters, and most of
these methods agree with integrated filter methods within 30%.
The differences observed between methods for OC and EC come largely from how OC and EC are
defined. They are defined on an operational basis, as there are no standard reference materials. Initial
efforts have been made to produce OC/EC reference materials at the National Institute of Science and
Technology (NIST) (Klouda et al., 2005; Lee, 2007). Klouda et al. (2005) described the development of
Reference Material 8785: Air Particulate Matter on Filter Media. Each reference filter is uniquely
identified by its air PM number and its gravimetrically determined mass of fine SRM 1649a, and each
filter has values assigned for total carbon, EC, and organic carbon mass fractions measured according to
both IMPROVE and NIOSH protocols. Lee et al. (2007) reported a method to create a reference filter
with a known amount of OC (as potassium hydrogen phthalate), and EC (as carbon black hydrosol).
Measurement methods for WSOC have been been developed recently (Greenwald et al., 2007;
Miyazaki et al., 2006; 2007; Sullivan et al., 2004; Sullivan et al., 2006; 2006; Yu et al., 2004). WSOC can
be measured on integrated filter samples, or in-situ measurement can be conducted by coupling with the
PILS-IC (Sullivan et al., 2004). For integrated filter samples, filters are extracted with deionized water
and followed by oxidation of total WSOC to C02. C02 can then be detected by either infrared
spectroscopy (IR) (Decesari et al., 2000; Kiss et al., 2002; Yang et al., 2003b), or FID (Yang et al.,
2003b), or pyrolysis gas chromatography/mass spectrometry (GC/MS) (Gelencser et al., 2000). A
correlation coefficient of 0.84 was reported by Sullivan et al. (2004) between in-situ and filter based
measurement of WSOC.
Further development and evaluation has been conducted on the measurement of BC with light
absorption instruments (Andreae and Gelencser, 2006; Arnott et al., 2003; Bae et al., 2004; Borak et al.,
2003; Cyrys et al., 2003; Hitzenberger et al., 1999; Kurniawan and Schmidt-Ott, 2006; Park et al., 2006a;
Saathoff et al., 2003; Sadezky et al., 2005; Slowik, 2007; Taha, 2007; Virkkula et al., 2007; Wallace,
December 2008
3-22
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
2000; Weingartner et al., 2003; Williams et al., 2006b; Wu et al., 2005). These instruments include the
aethalometer, particle absorption photometer, and photoacoustic analyzer. However, these instruments are
subject to interferences by particle scattering, interactions with the filter substrate, particle loading on
filters, and other pollutants (e.g. N02). Uncertainties of up to 50% were observed in the studies mentioned
above by comparing these methods and integrated filter methods.
Huang et al. (2006) reported the measurement of a stable isotope, 13C, in OC and EC with a thermal
optical transmission analyzer coupled with gas chromatography-isotope ratio mass spectrometer
(TOT-GC-IRMS). The ratio of 13C/12C in OC and EC can provide useful information on OC/EC source
categories and origin. The method was applied to Pacific2001 aerosol samples from the Greater
Vancouver area in Canada and produced a precision of -0.03%.
Organic Speciation
Organic matter makes up a substantial fraction of PM in all regions of the U.S. (U.S. EPA, 2004),
and 10 to 40% of the total organic matter is currently quantifiable at the individual compound level
(Poschl, 2005). Recent advancements in traditional solvent extraction gas chromatography/mass
spectrometry (GC/MS) and high pressure liquid chromatography (HPLC) as well as application of
thermal desorption (TD) techniques are helping to expand our understanding of the composition of
organic matter as well as improving detection limits for quantification of organic molecular marker
(OMM) compounds (Robinson et al., 2006; Schnelle-Kreis et al., 2005; Sheesley, 2007; Shrivastava et al.,
2007). In addition, information about organic functional groups can be obtained with Fourier transform
infrared spectrometry (FTIR) (Tsai and Kuo, 2006).
Recent advancements in GC/MS technology including inert electron ionization sources and
improved instrument sensitivity and scan rates for better OMM quantification, have increased its
application in organic aerosol characterization studies (Cass, 1998; Fraser et al., 2003; Graham et al.,
2003; Hays et al., 2002; Robinson et al., 2006; Schauer et al., 1996; Sheesley, 2007; Subramanian et al.,
2006; Watson et al., 1998; Zheng et al., 2002; 2006). Incorporation of high volume injection using
programmable temperature vaporization (PTV) (Engewald et al., 1999) has further increased detection
limits for trace level OMM compounds. High volume injection has the added benefit of preventing the
loss of semivolatile compounds (Swartz et al., 2003), and has been applied for analysis of PAHs using
low volume samplers (down to 5 Lpm), allowing for smaller required mass loadings (Bruno et al., 2007;
Crimmins and Baker, 2006). Since last review, HPLC analysis with fluorescence detection has also been
used frequently for quantification of semivolatile organic compounds in both the particle and gas phase
(Albinet et al., 2007; Barreto et al., 2007; Eiguren-Fernandez et al., 2003; Goriaux et al., 2006;
December 2008
3-23
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Murahashi, 2003; Ryno et al., 2006; Schauer et al., 2003; Stracquadanio et al., 2005; Temime-Roussel et
al., 2004a, b). Lengthy extraction and analysis times remain a limiting factor for these methods.
TD techniques bypass one of the time consuming steps in traditional solvent extraction analysis for
nonpolar organic compounds (n-alkanes, branched alkanes, cyclohexanes, hopanes, steranes, alkenes,
phthalates and PAHs). This is achieved by vaporizing and analyzing organic constituents directly from the
collection substrate, thereby bypassing the extraction step (Chow, 2007). Methods exist for both off-line
TD analysis of previously collected filter samples and semi-continuous TD analysis. Annex A is adapted
from Chow et al. (2007) and summarizes recent TD-GC/MS studies. The most common off-line method is
thermal desorption-gas chromatography/mass spectrometry (TD-GC/MS) (Hays and Lavrich, 2007).
Continuous or semi-continuous methods have been developed for direct analysis of individual organic
constituents by coupling TD with various forms of mass spectrometry (Smith et al., 2004; Tobias and
Ziemann, 1999; 2000; Voisin et al., 2003; Williams et al., 2006a). A comparison of measurement and
analytical specifications for filter analysis using solvent extraction and TD methods for organic speciation
are summarized in Annex A.
3.4.1.3. Ultrafine PM and PM Size Distribution
Instruments for measuring ultrafine PM developed during the past decade permit measurement of
size distributions of particles down to 3 nm in diameter with mobility particle sizers. Concentrations
down to this size range can be obtained by a Micro-Orifice Uniform Deposit Impactor (MOUDI). The
recently developed low pressure-drop ultrafine particle impactor coupled with a Beta Attenuation Monitor
(nano-BAM) can also provide ultrafine PM mass concentrations (Chakrabarti et al., 2004). A high
correlation coefficient was observed between MOUDIs and nano-BAMs, with a correlation of 0.96. A
50% cutpoint (d50) of 13-200 nm can be achieved by a high-volume slot-type ultrafine PM virtual
impactor (Middha and Wexler, 2006).
Methods are also being developed to measure the surface area of ultrafine particles. Wilson et al.
(2007b) suggested that the electrical aerosol detector (EAD) measurement might be a useful indicator of
the particle surface area deposited in the lung. This method can be potentially useful for examining the
association between health effects and particle surface areas.
Developments involving the condensation particle counter include use of de-ionized water as a
condensation media in lieu of butanol or n-propanol in condensation particle counters (Hering et al.,
2005; Hermann et al., 2007; Petaja, 2006). This development makes the condensation particle counter
(CPC) easier to use in field studies because water does not have some of the same chemical properties
(with respect to hazard and odor) as butanol or n-propanol. The performance of this CPC was reported to
be similar to the conventional butanol based CPC (Hering et al., 2005). Use of a battery of water and
December 2008
3-24
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
butanol-based CPCs was demonstrated to detect a range of solubilities in nucleation-mode particles
(Kulmala et al., 2007). Additionally, CPCs have been used to measure particles in the smaller end of the
ultrafine scale through adjustment of CPC cut-off diameters through tuning the temperature difference
between the CPC saturator and condenser (Kulmala et al., 2007) and improved charge reduction
techniques (Winkler et al., 2008). The latter method was effective in reducing the size of particles
detected by a CPC to less than 2 nm. These studies include assessment of errors related to these
developments with the CPC and generally show that counting efficiencies with these devices is upwards
of 95% (Hermann et al., 2007). Additionally, recent advancements have been made in development of fast
scanning methods for ultrafine particle size distributions, including diffusion screens (DS) (Feldpausch et
al., 2006) and fast integrated mobility scanners (FIMS) (Olfert et al., 2008).
It is also important to characterize the particle size distribution. For particles >0.1 (.im, several
instruments (e.g. DRUM, MOUDIs, and aerodynamic particle sizer [APS]) are available to measure
mass-based or count-based particle size distribution. For example, a PM monitoring system
(Aerodynamic Particle Sizer) with very sharp cut points between 0.1 and 10 (jm is now available (Peters,
2006; Zeng, 2006). For particles in this range, inertial forces are used to separate particles based on
impaction. For particles <0.1 |_im. particles can be separated by their electrical mobility, and as a result,
electrical mobility diameter, which is often used to describe ultrafine PM size distribution in lieu of
aerodynamic diameter. It has been necessary to develop techniques to change mobility diameters,
measured by the scanning mobility particle sizer (SMPS), to aerodynamic diameters, measured by the
APS, or vice versa, in order to merge the distributions spanning the ultrafine, accumulation, and coarse
modes. A variety of techniques for combining SMPS and APS diameters have been reported in the
literature (Hand and Kreidenweis, 2002; Khlystov et al., 2004; Morawska et al., 1999; 2007; Shen et al.,
2002; TSI., 2005). However, each of these techniques incurs some error of which the user must be aware.
3.4.1.4. Multiple-Component Measurements on Individual Particles
The 2004 PM AQCD discussed the aerosol time-of-flight mass spectrometry (ATOFMS). Recently,
the ATOFMS and several other aerosol mass spectrometry methods have been further developed. Both lab
and field comparisons have been conducted to evaluate the reliability of these types of instruments.
There are four types of commonly used aerosol mass spectrometry: (1) particle analysis by laser
MS (PALMS; National Oceanic and Atmospheric Administration [NOAA]); (2) rapid single particle mass
spectrometer (RSMS; University of Delaware); (3) aerosol time-of-flight MS (ATOFMS; TSI, Inc.); and
(4) AMS (Aerodyne) (Chow et al., 2008; Nash et al., 2006). The differences between these instruments
primarily come from the particle sizing methods of mass spectrometers, as shown in Annex A. Although
the technique varies, the underlying principle is to fragment each particle into ions, using either a
December 2008
3-25
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
high-power laser or a heated surface, and then a mass spectrometer to measure the mass to charge ratio of
each ion fragment in a vacuum.
These instruments were evaluated at the Atlanta, Houston, Fresno, Pittsburgh, New York, and
Baltimore supersites (Bein et al., 2005; Drewnick et al., 2004a; 2004b; Hogrefe et al., 2004; Jimenez et
al., 2003; Lake et al., 2003, 2004; Middlebrook et al., 2003; Phares et al., 2003; Qin and Prather, 2006;
Wenzel et al., 2003; Zhang et al., 2008). Measurements of the gross composition and abundance of
particles by these instruments were generally semi-quantitative, with the exception of AMS. Particles of
similar composition (e.g. 0C/S042~, Na/K/S042 . soot/hydrocarbon, and mineral particle types) were
characterized by these instruments during the studies mentioned above. N03 and S042 concentrations
measured with AMS were comparable with other continuous and filter-based methods, as mentioned in
Section 3.4.1.2. In addition, concentrations of different particle types can be obtained by the co-location
of these aerosol mass spectrometers and other particle sizing instruments, such as particle counters or the
MOUDI.
3.4.1.5. Emerging Methods
Alternative methods for measuring PM concentrations have been developed or improved since
2004. These include passive sampler development (discussed in Section 3.7.2.1) and satellite inference
techniques. Satellite remote-sensing techniques have been employed to infer ground-level PM2 5 mass
concentrations (Gupta et al., 2006; Pelletier et al., 2007; Schafer et al., 2008). Although the satellite
measurements can provide global coverage and concentrations in remote areas, the spatial resolution of
these measurements is still relatively large. Other shortcomings relating to operating principle and
interferences also exist.
3.4.2. Ambient Network Design
3.4.2.1. Monitor Siting Requirements
Where SLAMS PMi0 and PM2 5 monitoring is required, at least one of the sites must be a maximum
concentration site for that specific area. In 2007, there were 4,693 PMi0 monitors and 2,194 PM2 5
monitors reporting values to the EPA Air Quality System database (AQS). The AQS contains
measurements of air pollutant concentrations in the 50 states, plus the District of Columbia, Puerto Rico,
and the Virgin Islands, for the 6 criteria air pollutants and hazardous air pollutants. Criteria for siting
ambient monitors for PM at these sites are given in the CFR 40 Part 58 Appendix D, and
December 2008
3-26
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
SLAMS/NAMS/PAMS Network Review Guidance (U.S. EPA, 1998). The criteria described below are
summarized from CFR 40 Part 58 Appendix D.
The appropriate spatial scales for PM10 and PM2 5 SLAMS monitoring vary given differences in the
inertial behavior of coarse PM in contrast with fine. The scales at which networks are maintained are
listed below, and relevant scales for each size classification are provided in the subsections that follow.
¦ Micro (-5 - 100 m) and Middle (-100 - 500 m) scales—Some data uses associated with
microscale and middle scale measurements for PMi0 and PM2 5 include assessing the effects
of control strategies to reduce concentrations (especially for the 24-h averaging times), and
monitoring air pollution episodes.
¦ Neighborhood scale (-500 m - 4 km)—This scale applies where there is a need to collect air
quality data as part of an ongoing PMi0 and PM2 5 stationary source impact investigation.
Typical locations might include suburban areas adjacent to PMi0 and PM25 stationary
sources, for example, or for determining background concentrations as part of studies of
population responses to PM exposure.
¦ Urban scale (-4 - 50 km)—This scale applies for assessment of air quality at an urban scale,
although any given sampler is probably not representative of air quality as a whole across the
urban scale.
¦ Regional scale (-50 - 100s km)—This scale defines a fairly homogeneous rural area without
large sources. It can also be used to examine inter-urban variability and transport of pollution
across regions of the country.
PM10
Cities with populations in excess of one million are required to have between 2 and 10 monitors
(depending on concentration) while cities with populations less than one million are required to have
between 0 and 8 monitors (40 CFA Part 58, Appendix D). Except some circumstances where microscale
(<100 m, for maximum PMi0 exposure) monitoring may be appropriate, the most important scales to
characterize the emissions of PMi0 effectively from both mobile and stationary sources are the middle
scales (for short-term public exposure) and neighborhood scales (for trends and compliance with
standards). PMi0 measurements are obtained at standard temperature and pressure across the
NAMS/SLAMS networks (40 CFR Part 58).
December 2008
3-27
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
PM2.5
Monitor requirements for PM2 5 based on city population are similar to those for PM10 above.
Continuous PM2 5 monitors must be operated in no fewer than one-half of the minimum required sites in
each area. Most PM2 5 monitoring in urban areas should be representative of a neighborhood scale (for
trends and compliance with standards). Urban or regional scale sites are built to characterize regional
transport of PM2 5. In certain instances where population-oriented micro- or middle-scale PM2 5
monitoring are determined by the Regional Administrator to represent many such locations throughout a
metropolitan area, these smaller scales can be considered to represent community-wide air quality. PM2 5
measurements are obtained at the NAMS/SLAMS networks (40 CFR Part 58).
PM Species are monitored at both mostly urban (CSN) and mostly rural (IMPROVE) locations.
PM2 5 chemical speciation monitoring is currently conducted at 197 CSN sites
(http ://www.epa. gov/ttn/amtic/specgen.html). Within the CSN network, 53 locations are recognized as the
Speciation Trends Network (STN) operating on a sample schedule of one in every three days, while the
rest of the CSN typically operates very sixth day.
PM10-2.5
PM10-2.5 is not required to be monitored at SLAMS sites, but is required at NCore Stations. Middle
and neighborhood scale measurements are the most important station classifications for PMi0-2.5 to assess
the variation in coarse particle concentrations that would be expected across populated areas that are in
proximity to large emissions sources. PM10-2.5 chemical speciation monitoring and analyses is required at
NCore sites. Probing and monitoring path siting criteria for any specific monitoring site are given in CFR
40 Part 58 Appendix E, including horizontal and vertical placement, spacing from minor source, spacing
from obstructions, spacing from trees, and spacing from roadways. PMi0 measurements are converted to
local conditions to facilitate calculation of the difference between PMi0 and PM2 5.
3.4.2.2. Spatial and Temporal Coverage
Locations of PM2.5 and PM10 Monitors in Selected Metropolitan Areas in the U.S.
Fifteen metropolitan regions were chosen for closer investigation of monitor siting based on their
relevance to health studies analyzed in subsequent chapters of this ISA. These regions were: Atlanta,
Birmingham, Boston, Chicago, Denver, Detroit, Houston, Los Angeles, New York City, Philadelphia,
December 2008
3-28
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Phoenix, Pittsburgh, Riverside, Seattle, and St. Louis. Core-Based Statistical Areas (CBSAs) and
Combined Statistical Areas (CSAs), as defined by the U.S. Census Bureau (http://www.census/gov/).
were used to determine which counties, and hence monitors, to include for each metropolitan region.1
Appendix A displays PM2 5 and PMi0 monitor density with respect to population density for the fifteen
metropolitan regions. As an example, Figures 3-4 and 3-5 display PMi0 and PM25 monitor density with
respect to population density for Boston.
Tables 3-3 to 3-10 break down the population density around PM2 5 and PMi0 monitors for the total
population and for sub-populations of children aged 0-4 and 5-17 and elderly adults aged 65 and over for
each CSA/CBSA. Variation in percentage within a certain radius of the monitor was generally fairly low
for each city across the total population and child age groups. In the cases of Boston, Phoenix, Riverside,
and St. Louis for PM2 5 and Atlanta, Denver, and Riverside for PMi0, the elderly population's distribution
around the samplers varied more from the other groups. Between-city disparities in population density
were larger. For PM2 5, Denver (84%) and Los Angeles (82%) had the largest proportion of the total
population within 15 km of a monitor. Houston (30%) had the least population coverage with their PM2 5
monitors. For PMi0, Phoenix (89%) had the largest proportion of the total population within 15 km of a
monitor. Detroit (28%), Boston (36%), Seattle (37%), and Philadelphia (38%) had the smallest proportion
of the population within 15 km of a PM10 monitor. Proximity to monitoring stations is considered further
in sections 3.5 and 3.7 regarding spatial variability within cities. Figure 3-5 shows that the PM2 5 network
more closely samples near population centers compared with the PM10 network shown in Figure 3-4 for
Boston, although both PM10 and PM2 5 networks place at least one monitor in the city center.
1 A CBSA represents a county-based region surrounding an urban center of at least 10,000 people determined using 2000 census data and
replaces the older Metropolitan Statistical Area (MSA) definition from 1990. The CSA represents an aggregate of adjacent CBSAs tied by
specific commuting behaviors. The broader CSA definition was used when selecting monitors for the cities listed above with the exception of
Los Angeles, Riverside and Phoenix. Los Angeles and Riverside are contained within the same CSA, so the smaller CBSA definition was used
to delineate these two cities. Phoenix is not contained within a CSA, so the smaller CBSA definition was used for this city as well.
December 2008
3-29
DRAFT-DO NOT QUOTE OR CITE
-------
i Kilometers
0 15 30 60 90 120
Figure 3-4. PM10 monitor distribution in comparison with population density, Boston CSA.
2005 Population Density
PM 10 Monitors (5 km buffer)
Population per Sq Mile
| 0-650
| 651 - 1299
1300 - 6495
6496 - 12990
12991 - 32475
I 32476 - 129900
i Kilometers
10 20 30 40
December 2008 3-30 DRAFT-DO NOT QUOTE OR CITE
-------
Kilometers
2005 Population Density
PM 2 5 Monitors (5 km buffer)
Population per Sq Mile
| 0 -650
¦ 651 - 1299
12991 - 32475
32476 - 129900
Kilometers
0 15 30 60 90 120
Figure 3-5. PM2.5 monitor distribution in comparison with population density, Boston CSA.
December 2008
3-31
DRAFT-DO NOT QUOTE OR CITE
-------
Table 3-3. Proximity to PM2.5 monitors for the total population by city.
Proximity to PM10 Monitors
Region
Total CSA/CBSA
<1 km
<5 km
<10 km
< 15 km
N
N
%
N
%
N
%
N
%
Atlanta
4980447
30973
0.62
416440
8.36
1090497
21.90
1837983
36.90
Birmingham
1087703
23943
2.20
251309
23.10
473052
43.49
638471
58.70
Boston
4453936
21432
0.48
459019
10.31
1089852
24.47
1610722
36.16
Chicago
9537620
55646
0.58
844705
8.86
2374966
24.90
3843908
40.30
Denver
2379716
20401
0.86
330202
13.88
899951
37.82
1440920
60.55
Detroit
4590501
14050
0.31
309619
6.74
748977
16.32
1300705
28.33
Houston
5398706
36795
0.68
832767
15.43
2227314
41.26
3141028
58.18
Los Angeles
13057686
52054
0.40
1404370
10.76
4899179
37.52
9075754
69.51
New York
rJa
n la
n/a
n la
n/a
n la
n/a
rJa
n la
Philadelphia
5848871
23988
0.41
376968
6.45
1091530
18.66
2238306
38.27
Phoenix
3818147
99520
2.61
1255430
32.88
2615738
68.51
3416682
89.49
Pittsburgh
2419053
61291
2.53
667265
27.58
1234171
51.02
1618336
66.90
Riverside
3781063
61356
1.62
895615
23.69
2360272
62.42
2922799
77.30
Seattle
3246877
4771
0.15
219722
6.77
709070
21.84
1205927
37.14
St. Louis
2810628
27872
0.99
380411
13.53
891695
31.73
1212543
43.14
Percentages are given with respect to the total population per city provided.
Table 3-4. Proximity to PM10 monitors for the total population by city.
Proximity to PM10 Monitors
Region
Total CSA/CBSA
<1 km
<5 km
<10 km
< 15 km
N
N
%
N
%
N
%
N
%
Atlanta
4980447
30973
0.62
416440
8.36
1090497
21.90
1837983
36.90
Birmingham
1087703
23943
2.20
251309
23.10
473052
43.49
638471
58.70
Boston
4453936
21432
0.48
459019
10.31
1089852
24.47
1610722
36.16
Chicago
9537620
55646
0.58
844705
8.86
2374966
24.90
3843908
40.30
Denver
2379716
20401
0.86
330202
13.88
899951
37.82
1440920
60.55
Detroit
4590501
14050
0.31
309619
6.74
748977
16.32
1300705
28.33
Houston
5398706
36795
0.68
832767
15.43
2227314
41.26
3141028
58.18
Los Angeles
13057686
52054
0.40
1404370
10.76
4899179
37.52
9075754
69.51
New York
rJa
n la
n/a
n la
n/a
n la
n/a
rJa
n la
Philadelphia
5848871
23988
0.41
376968
6.45
1091530
18.66
2238306
38.27
Phoenix
3818147
99520
2.61
1255430
32.88
2615738
68.51
3416682
89.49
December 2008
3-32
DRAFT-DO NOT QUOTE OR CITE
-------
Pittsburgh
2419053
61291
2.53
667265
27.58
1234171
51.02
1618336
66.90
Riverside
3781063
61356
1.62
895615
23.69
2360272
62.42
2922799
77.30
Seattle
3246877
4771
0.15
219722
6.77
709070
21.84
1205927
37.14
St. Louis
2810628
27872
0.99
380411
13.53
891695
31.73
1212543
43.14
Percentages are given with respect to the total population per city provided.
Table 3-5.
Proximity to PM2.5 monitors for children aged 0-4 by city.
Proximity to PM2.5 Monitors
Region
Total CSA/CBSA
<1 km
<5 km
<10 km
< 15 km
N
N
%
N
%
N
%
N
%
Atlanta
317949
1263
0.40
34671
10.90
122351
38.48
197996
62.27
Birmingham
70482
884
1.25
15768
22.37
44086
62.55
55224
78.35
Boston
277628
6245
2.25
66920
24.10
123763
44.58
177769
64.03
Chicago
675274
14081
2.09
234890
34.78
465860
68.99
580204
85.92
Denver
153531
1066
0.69
24149
15.73
80140
52.20
117083
76.26
Detroit
312177
3817
1.22
74109
23.74
169820
54.40
223174
71.49
Houston
379467
972
0.26
18781
4.95
77667
20.47
125829
33.16
Los Angeles
953522
9187
0.96
204772
21.48
583222
61.17
820231
86.02
New York
1243377
48129
3.87
471327
37.91
787795
63.36
955967
76.88
Philadelphia
369018
7285
1.97
125606
34.04
223485
60.56
275691
74.71
Phoenix
254040
3314
1.30
41675
16.40
89794
35.35
128482
50.58
Pittsburgh
134859
2260
1.68
31406
23.29
72557
53.80
102891
76.30
Riverside
264700
3785
1.43
60346
22.80
144522
54.60
175659
66.36
Seattle
198596
942
0.47
19252
9.69
57459
28.93
92985
46.82
St. Louis
179092
2695
1.50
36535
20.40
86315
48.20
112672
62.91
Percentages are given with respect to the total population per city provided.
December 2008
3-33
DRAFT-DO NOT QUOTE OR CITE
-------
Table 3-6. Proximity to PM10 monitors for children aged 0-4 by city.
Proximity to PM10 Monitors
Region
Total CSA/CBSA
<1 km
<5 km
<10 km
< 15 km
N
N
%
N
%
N
%
N
%
Atlanta
317949
1465
0.46
23703
7.45
66981
21.07
117010
36.80
Birmingham
70482
1675
2.38
16691
23.68
31553
44.77
41457
58.82
Boston
277628
1530
0.55
22486
8.10
57300
20.64
87595
31.55
Chicago
675274
4259
0.63
61898
9.17
177799
26.33
293181
43.42
Denver
153531
786
0.51
22534
14.68
60633
39.49
94729
61.70
Detroit
312177
1260
0.40
24363
7.80
56992
18.26
98211
31.46
Houston
379467
3109
0.82
67631
17.82
173298
45.67
235197
61.98
Los Angeles
953522
3740
0.39
112759
11.83
384078
40.28
693715
72.75
New York
n la
n/a
rJa
n la
rJa
n/a
rJa
rJa
n la
Philadelphia
369018
1821
0.49
24351
6.60
65371
17.71
133324
36.13
Phoenix
254040
8450
3.33
95136
37.45
184308
72.55
232869
91.67
Pittsburgh
134859
3239
2.40
36237
26.87
67452
50.02
89153
66.11
Riverside
264700
4502
1.70
71812
27.13
178212
67.33
212080
80.12
Seattle
198596
378
0.19
13916
7.01
40826
20.56
68559
34.52
St. Louis
179092
2266
1.27
26457
14.77
61592
34.39
79845
44.58
Percentages are given with respect to the total population per city provided.
Table 3-7. Proximity to PM2.5 monitors for children aged 5-17 by city.
Proximity to PM2.5 Monitors
Region
Total CSA/CBSA
<1 km
<5 km
<10 km
< 15 km
N
N
%
N
%
N
%
N
%
Atlanta
813107
2739
0.34
78623
9.67
297032
36.53
492499
60.57
Birmingham
192830
2404
1.25
44556
23.11
119723
62.09
151012
78.31
Boston
748858
15347
2.05
176638
23.59
329170
43.96
475482
63.49
Chicago
1772017
35584
2.01
602396
33.99
1195980
67.49
1505243
84.95
Denver
398461
2505
0.63
60214
15.11
196983
49.44
293694
73.71
Detroit
869389
11105
1.28
207420
23.86
474076
54.53
615848
70.84
Houston
988206
2464
0.25
46060
4.66
187215
18.94
309878
31.36
Los Angeles
2482440
23781
0.96
516653
20.81
1474428
59.39
2118897
85.36
New York
3266360
117638
3.60
1200357
36.75
2043076
62.55
2480535
75.94
Philadelphia
1074283
20096
1.87
361257
33.63
652630
60.75
807584
75.17
Phoenix
619044
7037
1.14
92828
15.00
199177
32.17
302349
48.84
December 2008
3-34
DRAFT-DO NOT QUOTE OR CITE
-------
Pittsburgh
406762
6777
1.67
90122
22.16
214227
52.67
309488
76.09
Riverside
756027
9182
1.21
160971
21.29
398432
52.70
495123
65.49
Seattle
548642
2556
0.47
52257
9.52
155593
28.36
253031
46.12
St. Louis
528319
7394
1.40
103816
19.65
250730
47.46
331111
62.67
Percentages are given with respect to the total population per city provided.
Table 3-8.
Proximity to PM10 monitors for children aged 5-17 by city.
Proximity to PM10 Monitors
Region
Total CSA/CBSA
<1 km
<5 km
<10 km
< 15 km
N
N
%
N
%
N
%
N
%
Atlanta
813107
3064
0.38
51765
6.37
155239
19.09
280891
34.55
Birmingham
192830
4745
2.46
47958
24.87
86186
44.70
114890
59.58
Boston
748858
4182
0.56
61075
8.16
148703
19.86
226634
30.26
Chicago
1772017
11096
0.63
165032
9.31
473575
26.73
783967
44.24
Denver
398461
1551
0.39
50915
12.78
139990
35.13
228485
57.34
Detroit
869389
3126
0.36
61201
7.04
150762
17.34
269196
30.96
Houston
988206
6960
0.70
149188
15.10
406773
41.16
567427
57.42
Los Angeles
2482440
9169
0.37
285327
11.49
968855
39.03
1801841
72.58
New York
n la
n la
n/a
n/a
rJa
n/a
n/a
n/a
n la
Philadelphia
1074283
4848
0.45
68349
6.36
189689
17.66
388207
36.14
Phoenix
619044
18456
2.98
208231
33.64
425518
68.74
562374
90.85
Pittsburgh
406762
9562
2.35
103569
25.46
198260
48.74
267091
65.66
Riverside
756027
11877
1.57
193891
25.65
492092
65.09
601629
79.58
Seattle
548642
729
0.13
34323
6.26
106779
19.46
178574
32.55
St. Louis
528319
6593
1.25
80897
15.31
180583
34.18
232835
44.07
Percentages are given with respect to the total population per city provided.
December 2008
3-35
DRAFT-DO NOT QUOTE OR CITE
-------
Table 3-9. Proximity to PM2.5 monitors for adults aged 65 and older by city.
Proximity to PM2.5 Monitors
Region
Total CSA/CBSA
<1 km
<5 km
<10 km
< 15 km
N
N
%
N
%
N
%
N
%
Atlanta
326656
1400
0.43
33070
10.12
127765
39.11
194901
59.67
Birmingham
134563
1618
1.20
29954
22.26
84200
62.57
106389
79.06
Boston
550114
10919
1.98
127385
23.16
266527
48.45
380376
69.14
Chicago
990352
17116
1.73
339558
34.29
700994
70.78
869475
87.79
Denver
93460
1847
0.95
38143
19.72
113825
58.84
163801
84.67
Detroit
532845
4983
0.94
123179
23.12
307678
57.74
408426
76.65
Houston
366637
1010
0.28
14911
4.07
66728
18.20
117629
32.08
Los Angeles
1206715
9653
0.80
229893
19.05
688833
57.08
984878
81.62
New York
2306151
76951
3.34
814370
35.31
1422463
61.68
1779594
77.17
Philadelphia
758833
13323
1.76
251459
33.14
487003
64.18
605663
79.82
Phoenix
388150
2738
0.71
39833
10.26
90304
23.27
142084
36.61
Pittsburgh
430748
8933
2.07
111050
25.78
249269
57.87
345281
80.16
Riverside
342334
3024
0.88
50901
14.87
129836
37.93
170933
49.93
Seattle
308746
1721
0.56
29426
9.53
101223
32.79
156484
50.68
St. Louis
350324
5401
1.54
83528
23.84
192532
54.96
244929
69.92
Percentages are given with respect to the total population per city provided.
Table 3-10. Proximity to PM10 monitors for adults aged 65 and older by city.
Proximity to PM10 Monitors
Region
Total CSA/CBSA
<1 km
<5 km
<10 km
< 15 km
N
N
%
N
%
N
%
N
%
Atlanta
326656
2115
0.65
35448
10.85
93903
28.75
139240
42.63
Birmingham
134563
3663
2.72
35628
26.48
66839
49.67
86299
64.13
Boston
550114
1983
0.36
43421
7.89
125579
22.83
205946
37.44
Chicago
990352
7619
0.77
107539
10.86
291704
29.45
441729
44.60
Denver
193460
2100
1.09
28705
14.84
88653
45.83
143801
74.33
Detroit
532845
1555
0.29
41832
7.85
99682
18.71
167716
31.48
Houston
366637
2085
0.57
57413
15.66
166715
45.47
219603
59.90
Los Angeles
1206715
4693
0.39
126694
10.50
422716
35.03
810068
67.13
New York
n la
n/a
n/a
n/a
n/a
n/a
rJa
rJa
n/a
Philadelphia
758833
2740
0.36
49413
6.51
154535
20.36
322700
42.53
Phoenix
388150
8605
2.22
119306
30.74
267456
68.91
348464
89.78
December 2008
3-36
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Pittsburgh 430748 12222 2.84 124874 28.99 231658 53.78 297674 69.11
Riverside 342334 4181 1.22 65499 19.13 182615 53.34 236900 69.20
Seattle 308746 498 0.16 22283 7.22 72929 23.62 122424 39.65
St. Louis 350324 4316 1.23 55833 15.94 117743 33.61 172535 49.25
Percentages are given with respect to the total population per city provided.
Annex A shows the locations of PM2.5 and PMi0 monitors for the 15 CSAs/CBSAs in relation to
major roadways, including Interstate highways, U.S. highways, state highways, and other major roadways
required for network connectivity. In most cases, the monitors are concentrated near the center of the
CSA/CBSA. Regional background sites are not included on the maps unless they lie within the
CSA/CBSA.
Comparison of Monitors at Supersites
Annex A lists parameters and findings for various supersite mass monitoring efforts, with
methodology for the various sites listed in this table along with results. Inter-sampler comparison results
varied widely. Conditions impacting agreement between samplers included volatility of the particles and
related operation temperature of the monitor, particle shape, and hygroscopicity of the particles. Particle
agglomeration may also impact results for laser photometric methods. Annex A compares supersite
monitoring results for particle-bound water, N03 , S042 . carbon, and mass spectrometry, respectively.
Example results are highlighted here; more detail can be found in the tables. For particle-bound water,
variability in the results was reported at 20-43% (Kidwell and Ondov, 2004). Khlystov et al. (2005) and
Stanier et al. (2007) reported growth factors of 2-14%, and measurement agreement varied from 20-35%
(Solomon et al., 2003; Weber et al., 2008). For S042 , Solomon (2003) reported greater than 50%
variability of analyzed S042 compared with a reference filter, although Weber et al. (2008) and Zhang et
al. (2005b) reported agreement within 16%. Drewnick et al. (2003) and Hogrefe et al. (2004) reported
25% inlet losses but only 2-3% loss for the AMS. For carbon, Solomon et al. (2003) showed that that OC
measured with denuded samplers was lower than reference, while non-denuded samples had higher OC.
Subramanian et al. showed a positive artifact when QBT was used. At the same time, they measured a
negative artifact of approximately 6.3% of particulate OC. Bae et al. (2004) noted denuder breakthrough
of only 5%. Solomon et al. (2003) showed EC agreement to vary from 20-200% of the reference filter.
The variable responses reported in these studies reflect the fact that sampler operating conditions and
properties of the aerosol (i.e. size, chemical composition, volatility, shape) are also highly variable and
present challenges for identification of errors and interpretation of results.
December 2008
3-37
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
3.5. Ambient PM Concentrations
This section describes measurements of ambient PM and its components made since the 2004 PM
AQCD (U.S. EPA, 2004) and recent advancements in understanding of the spatiotemporal concentration
distribution of PM and its constituents. Emphasis is placed on the period from 2005-2007, which
incorporates the most recent, validated EPA Air Quality System (AQS) data available at the time this
document was prepared.
When the 2004 PM AQCD was written, the full nationwide PM2 5 compliance monitoring network
had only recently been deployed, providing three years (1999 to 2001) of measurements. For this PM
ISA, the network has been active for eight or nine years depending on location. Therefore, this document
contains substantial new data for examining the spatiotemporal distribution of PM2 5. Furthermore, by
selecting locations where PM10 and PM2 5 measurements are co-located, further information about the
spatiotemporal distribution of the PM10-25 size fraction has been gained since the last AQCD. A large
amount of new information has emerged over the last several years regarding PM2 5 composition and
ultrafine particle concentrations. Compliance monitoring does not apply for ultrafine particles because
there is no ambient standard for them.
Spatial distributions of PM across a range of geographic scales are covered in Section 3.5.1.
Temporal behavior including trends, seasonality and hourly variability are covered in Section 3.5.2.
Statistical associations between different size fractions of PM and copollutants including CO, N02, 03
and S02 are included in Section 3.5.3. Finally, source attribution methodologies, results and uncertainties
are covered in Section 3.5.4.
Unless otherwise specified, the PM data utilized throughout this section comes from the AQS.
Monitors reporting to the AQS are not uniformly distributed across the U.S. Monitors are far more
abundant in urban areas than rural ones, so true rural spatiotemporal distributions may differ from those
reported here. Furthermore, biases exist for some PM constituents (and hence total mass) owing to
volatilization losses of N03 and other semi-volatile compounds and, conversely, retention of
particle-bound water with hygroscopic species. The magnitude of these effects is likely to be
region-specific. Such regional differences are addressed in the following section.
3.5.1. Spatial Distribution
Spatial scales of interest for PM include global and continental (>1000 km), regional (100 to 1000
km), urban (4 to 100 km) and neighborhood (<4 km) scales (Watson et al., 1997). Focusing primarily on
data available from the AQS, this section has been divided into three sub-sections based on spatial scale:
December 2008
3-38
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
variability across the U.S., urban-scale variability and neighborhood-scale variability. Variation in PM
concentration depends on the spatial scale and magnitude of PM sources, formation and removal
mechanisms, and transport and dispersion of PM. These different sources and processes can cause
substantial variation in particle size distribution and chemistry. Consequently, the following sections are
further subdivided to the extent possible into PM size fractions and composition.
3.5.1.1. Variability across the U.S.
PM10
Figure 3-6 shows the 3-y mean of the 24-h PMi0 concentrations by county across the U.S. for
2005-2007. The data used in generating this map are from the AQS database after applying a
completeness criteria of 75 % per quarter (or 11/15 measurements for a one-in-six-day sampling
schedule). The highest annual averages for PMi0 (>51 (.ig/nr1) generally occur in inland southern
California and the populous counties of southern Arizona and central New Mexico. Of the 3,225 U.S.
counties, 676 (12%) contained PMi0 data meeting the completeness criteria in all three years (2005-2007)
and have been included in Figure 3-6. These 676 counties incorporate approximately 43% of the U.S.
population. The fraction of the population for each reported concentration level is also shown in the
figure. Given the number of counties with no data, the non-uniform spacing of the monitors, and the
population within each reporting county, this should only be taken as a rough estimate of the relationship
between population and ambient monitors and concentrations.
December 2008
3-39
DRAFT-DO NOT QUOTE OR CITE
-------
Concentration Range
B > 51 |ig/m3 [7 counties]
41-50 tig/m3 [15 counties]
31-40 |ig/m3 [378 counties]
"21-30 tig/m3 [162 counties]
a < 20 tig/m3 [114 counties]
~ No data
Population
(millions)
Figure 3-6. Average 24-h PM10 concentration by county derived from FRM or FEM monitors, 2005—
2007. The population bar shows the number of people residing within counties that
reported county-wide average values in each of the concentration ranges.
Table 3-11. PM10 distributions derived from AQS data (concentration in |jg/m3).
n
Mean
Percentiles
Max
1
5
10
25
50
75
90
95
99
2005-2007 PMn FOR DIFFERENT AVERAGING PERIODS
Annual avg * (24-h FRM and 1-h FEM)
674
25
9
14
16
19
22
28
35
43
58
81
24-h avg. (24-h FRM and 1-h FEM)
326,675
26
3
6
9
14
21
32
46
59
97
8299
24-h avg. (24-h FRM)
167,310
25
2
6
9
14
21
31
45
57
91
8299
24-h avg. (1-h FEM)
156,931
26
4
7
9
14
21
32
48
62
105
979
1-h avg. (1-h FEM)
3,767,533
27
1
4
6
11
19
32
51
69
145
8540
PMn ANNUAL AND SEASONAL STRATIFICATION USING 24-H AVG. FRM AND FEM DATA
2005
107,524
25
2
6
9
13
21
31
46
58
93
1441
2006
109,505
26
3
6
9
13
21
32
46
59
101
8299
2007
109,646
26
4
7
9
14
21
32
47
60
99
2253
Winter (December-February)
80,959
23
2
5
7
11
17
27
42
57
99
8299
Spring (March-May)
82,772
25
2
6
8
13
20
31
45
58
96
2253
Summer (June-August)
81,351
29
6
10
12
18
25
35
49
60
92
1839
Fall (September-November)
81,593
26
3
7
9
14
21
32
48
62
102
1212
December 2008
3-40
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
2005-2007 PMn IN INDIVIDUAL CSAS/CBSAS USING 24-H AVG. FRM AND FEM DATA
Atlanta
24
23 31
44
57
Birmingham
5,478
34
12
28 43
64
82 120
241
Boston
1,412
17
15 22 30 36
50
58
Chicago
6,165
26
23
32
45
55
78
214
Denver
4,706
10
12
25
35
47
54
75
118
Detroit
1,407
30
7 10 12
26 38 53 64
Houston
1,397
31
7
10
12
23
34
56
137
248
Los Angeles
2,020
27
25 33 42 51
74
New York
514
19
17
25
35
40
Philadelphia
4,207
19
7
17 24 34 40
52
Phoenix
12,005
52
7 14 19 29 44 65 91
112 166
2253
Pittsburgh
12,677
24
7
19
31
45
57
83
157
Riverside
4,327
35
30 45 64 75
111
1212
Seattle
2,136
19
7
17
23
31
37
52
79
St. Louis
2,464
33
10 12
42 59 74 114
315
All 15 CSAs/CBSAs
62,783
32
10
25
77
120
2253
Not in the 15 CSAs/CBSAs
24
20
30
43
54
"straight annual average without quarterly weighting
Table 3-11 contains summary statistics for PMi0 reported to AQS for the period 2005-2007. Both
FRM and FEM data are included in the table. The majority of observations are 24-h filter-based FRM
measurements, but numerous sites also report 1-h FEM data. To facilitate a distributional comparison
between 24-h and 1-h data reporting, the FRM and FEM data have been separated in Table 3-11. The
table also includes the data stratified by year (2005, 2006 and 2007) and season: winter
(December-February), spring (March-May), summer (June-August), and fall (September-November).
Fifteen CSAs/CBSAs were chosen for their importance in recent PM health studies, as described in
Section 3.4.
PM2.5
Figure 3-7 shows the 3-y mean of the 24-h PM25 concentrations by county across the U.S. for
2005-2007. These data are obtained from the same source as PMi0 with the same 11+ days per quarter
completeness criteria applied. The San Joaquin Valley and inland southern California reported high PM2 5
24-h concentrations above 20 |ig/nr\ In addition, Jefferson County containing Birmingham, AL and
Allegheny County containing Pittsburgh, PA show annual average PM2 5 concentrations above 18 (ig/m3.
Of the 3,225 U.S. counties, 540 (17%) had PM2 5 data meeting the completeness criteria in all three years
(2005-2007) and have been included in Figure 3-7. These 540 counties represent roughly 63% of the U.S.
population.
December 2008
3-41
DRAFT-DO NOT QUOTE OR CITE
-------
1 Table 3-12 contains summary statistics for PM2 5 reported to AQS for the period 2005-2007. All
2 24-h FRM and 1-h FRM-like1 data reported to AQS are included in the table. On a national basis using
3 the 2005-2007 data in Tables 3-11 and 3-12, the mean 24-h PM10 concentration (26 (.ig/nr1) is slightly
4 more than twice the mean 24-h PM2 5 concentration (12 |_ig/m3). Therefore, approximately half the PM
5 mass is in the fine mode and half in the coarse mode. The distribution between fine and coarse PM varies
6 substantially by location with a larger fraction of PM mass in the coarse mode in drier climates like
7 Phoenix and Denver and a larger fraction in the fine mode in cities like Pittsburgh and Philadelphia.
8 Comparisons of PM2 5 to PMi0 as reported to AQS should be used with caution, however, since PM2 5
9 concentrations are reported for local conditions while PMi0 concentrations are converted to STP before
10 reporting.
1 FRM-like refers to PM2 5 concentration data associated with the parameter code "88502 - Acceptable PM2.5 AQI and Speciation Mass" in the
EPA Air Quality System. These data were collected by continuous instruments which are not approved as FRM or FEM, and consequently EPA
does not use these data for regulatory purposes. These data are denoted as "FRM-like" because state and local monitoring agencies have
individually decided that the continuous instruments reporting these data have a degree of agreement with FRM/FEM methods that is sufficient
in their opinion for the data to be used in public advisories regarding current air quality. In some cases, these data include statistical adjustments
by the state/local monitoring agency based on one-time or ongoing correlation analysis with co-located FRM/FEM monitors, intended to
improve the "FRM-likeness" of the continuous concentration data (see, for example, (Bortnick et al., 2002). State/local monitoring agency
decisions about whether to adjust continuous PM2.5 data and whether their raw or adjusted continuous PM2.5 data should be associated with
parameter code 88502 were informed by non-binding EPA guidance issued in 2006 (Technical Note on Reporting PM2.5 Continuous Monitoring
and Speciation Data to the Air Quality System (AQS) http://www.epa. gov/ttn/amtic/files/ambient/pm25/datamang/contrept.pdf).
December 2008
3-42
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
Concentration Range
a > 20.1 |ig/m3 [1 county]
18.1 - 20.0 ^g/m3 [7 counties]
15.1 - 18.1 [ig/m3 [53 counties]
a 12.1 - 15.0 [ig/m3 [242 counties]
B £ 12.0 tig/m3 [237 counties]
~ No data
Population
(millions)
Figure 3-7. Average 24-h PM2.5 concentration by county derived from FRM or FRM-like data, 2005-
2007. The population bar shows the number of people residing within counties that
reported county-wide average values in each of the concentration ranges.
PM10-2.5
Co-located PM10 and PM2 5 measurements from the AQS network were used to generate the
PM10-2.5 concentrations by county across the U.S. for 2005-2007 in Figure 3-8. Only FRM or FRM-like
samplers were considered in calculating PM10_2 5 to avoid complications with vastly different sampling
protocols (e.g., flow rates) between the independent PMi0 and PM2 5 measurements. The PM2 5
concentrations are reported to AQS using local conditions; the PMi0 data were adjusted to local conditions
on a daily basis using temperature and pressure measurements from the nearest National Weather Service
station. Figure 3-8 has considerably less coverage than the corresponding PMi0 and PM2.5 figures as a
result of the monitor selection criteria and the fact that not all PMi0 and PM2 5 monitors in a given region
are co-located. The 40 counties included in Figure 3-8 incorporate less than 5% of the U.S. population. Of
the 3,225 U.S. counties, only 40 (1%) met the completeness and co-location criteria in all three years
(2005-2007) and therefore the available measurements do not provide sufficient information to assess
regional-scale coarse PM concentration distributions.
December 2008
3-43
DRAFT-DO NOT QUOTE OR CITE
-------
Table 3-12. PM2.5 distributions derived from AQS data (concentration in |jg/m3).
n
Percentiles
1
5
10
25
50
75
90
95
99
2005-2007 PMis FOR DIFFERENT AVERAGING PERIODS
Annual avg * (24-h FRM)
794
12
5
6
8
10
13
14
15
16
19
22
24-h avg. (24-h FRM)
349,028
12
2
4
4
7
10
16
23
28
39
193
24-h avg. (1-h FRM-like)
183,057
10
1
2
3
5
8
13
19
24
35
126
1-havg. (1-h FRM-like)
4,403,817
10
0
1
2
4
8
13
21
27
42
828
PMis ANNUAL AND SEASONAL STRATIFICATION USING 24-H AVG. FRM DATA
2005
114,346
13
2
4
5
7
11
17
24
30
42
133
2006
113,197
12
2
4
4
7
10
15
21
26
36
193
2007
121,485
12
2
4
4
7
10
16
22
27
40
145
Winter (December-February)
86,286
12
2
4
5
7
10
15
22
27
44
193
Spring (March-May)
88,489
11
2
3
4
6
9
14
20
24
33
145
Summer (June-August)
86,830
14
2
4
5
8
12
19
26
31
40
133
Fall (September-November)
87,423
12
2
3
4
6
10
15
22
26
39
126
2005-2007 PMis IN INDMDUAL CSAS/CBSAS USING 24-H AVG. FRM DATA
Atlanta
4,939
15
4
6
7
10
14
19
25
29
37
145
Birmingham
4,869
16
4
6
7
10
15
21
29
34
47
64
Boston
8,464
10
2
3
4
5
9
13
20
24
32
50
Chicago
10,308
14
3
4
6
8
13
18
25
31
42
65
Denver
4,192
9
2
3
4
6
8
10
14
18
31
61
Detroit
5,223
14
2
3
5
7
12
19
26
31
45
82
Flouston
1,342
15
4
6
8
10
14
18
23
26
34
44
Los Angeles
6,600
15
3
5
6
9
13
18
25
32
50
133
New York
15,826
13
2
4
4
6
10
17
24
29
39
58
Philadelphia
7,541
14
3
4
5
8
12
18
25
30
38
63
Phoenix
1,634
10
2
3
4
6
9
12
17
21
32
77
Pittsburgh
5,783
16
3
5
6
9
13
20
29
36
52
101
Riverside
2,751
17
3
5
6
10
14
21
31
40
58
106
Seattle
1,297
9
2
3
3
4
7
10
20
29
43
68
St. Louis
6,887
14
3
5
6
9
13
18
24
29
40
50
All 15 CSAs/CBSAs
87,656
14
2
4
5
7
12
17
25
30
42
145
Not in the 15 CSAs/CBSAs
261,372
12
2
3
4
6
10
15
22
27
38
193
*Straight annual average without quarterly weighting.
December 2008
3-44
DRAFT-DO NOT QUOTE OR CITE
-------
PM10 25 24-hour Average Concentration, 2005-2007
25#
201
15#
10#
50
Population
(millions) ^
Concentration Range
m >26 |ig/m3 [1 county]
"21-25 ^g/m3 [3 counties]
16-20 ^g/m3 [3 counties]
c 11-15 |ig/m3 [17 counties]
a < 10 pg/m3 [16 counties]
~ No data
Figure 3-8. Average 24-h PM10-2.5 concentration by county derived from co-located low volume
FRM PM10 and PM2.5 monitors, 2005-2007. The population bar shows the number of
people residing within counties that reported county-wide average values in each of
the concentration ranges.
1 Table 3-13 contains summary statistics for PM10.2 5 forthe period 2005-2007 similar to those above
2 for PM10 and PM2 5. Given the FRM requirement applied here for calculating PM10.25, no continuous data
3 were incorporated into this table. Using all available co-located PM measurements from 2005-2007, the
4 mean 24-h PM10-2.5 concentration (13 (.ig/nr1) is roughly equivalent to the mean 24-h PM2.5 concentration
5 (12 (ig/m3).
PM Constituents
6 Figures 3-9 through 3-13 contain U.S. concentration maps for OC, EC, S04, N03, and NH4 from
7 the CSN network for the period 2005-2007.
December 2008
3-45
DRAFT-DO NOT QUOTE OR CITE
-------
Table 3-13. PM10-2.5 distributions derived from AQS data (concentration in |jg/m3).
n
Mean
Percentiles
Max
1
5
10
25
50
75
90
95
99
2005-2007 PM10-2., FOR DIFFERENT AVERAGING PERIODS
Annual avg * (low volume FRM)
43
12
4
5
6
9
11
14
20
22
40
40
24-h avg. (low volume FRM)
12,027
13
-3
1
2
6
10
17
26
33
54
246
PM10-25 ANNUAL AND SEASONAL STRATIFICATION USING 24-h AVG. LOW VOLUME FRM DATA
2005
3,990
12
-5
0
2
5
10
16
26
33
52
246
2006
4,037
13
-2
1
2
6
10
17
27
34
56
182
2007
4,000
13
-2
1
3
6
11
18
26
33
56
148
Winter (December-February)
2,942
11
-5
-1
1
4
8
15
27
34
56
246
Spring (March-May)
3,088
13
-2
1
2
5
10
17
26
33
62
151
Summer (June-August)
2,968
14
-2
3
5
8
12
18
25
31
44
93
Fall (September-November)
3,029
14
-2
1
3
6
11
18
28
34
60
148
2005-2007 PMw-2. , IN INDIVIDUAL CSAS/CBSAS USING 24-H AVG. LOW VOLUME FRM DATA**
Atlanta
167
10
-4
1
2
5
9
13
18
21
30
46
Boston
340
7
-2
1
2
4
6
9
12
16
25
27
Chicago
161
5
-8
-4
-3
1
4
8
14
19
37
37
Denver
353
20
0
4
6
11
19
28
36
42
59
78
New York
338
9
-16
-2
1
5
8
12
17
23
34
56
Phoenix
163
22
-3
8
11
16
20
29
35
46
67
70
All 6 CSAs/CBSAs
1,522
12
-6
0
2
5
10
17
27
34
51
78
Not in the 6 CSAs/CBSAs
10,505
13
-2
1
2
6
10
17
26
33
56
246
*straight annual average without quarterly weighting
**no co-located FRM PM10 and FRM-like PM25 monitors present in Birmingham, Detroit, Houston, Los Angeles, Philadelphia, Pittsburgh, Riverside, Seattle or St. Louis.
December 2008
3-46
DRAFT-DO NOT QUOTE OR CITE
-------
OC - Annual
9 ,*
g7
1 .
X .
©° ,°o
o o o_
o
fij
>
%
o j O o^O .^
0 o
OC(ug/m3): <=1.25 >1.25 to 2.5
>2.5 to 3.75 >3.75-5
OC - Spring
<6
o o
• 2 ' »®s
® S ^0=?" «
o 125 to 2.5
• •
>375-5 >5
OC - Summer
$
o o
<3*
8„
o afT o»
o
G>
'1*5#*°
OC(ugftT»3) <=125 >12510 2 5
>2 5 to 3.75 >3 75-5
>5
:.-.
t.
\ .
OC - Fall
0 i V- \
« 0 *•*» °
*© *?*•>
¦ L • • S V ?»
o ® %_ °*
o o °
• o O • •
0C(uym3> «»12S >12SW3S >250 3 75 >375-5 >5
o ••
o •
w
I
OC - Winter
8 °
©
© o
!«°;
'CP
oe o
" 0
°qg 0f-«0 =
, <= 08o Tp
<® «•* 0
°8
„ ,»= O
0C(ug««3> <«12S >125W2S
O
>26103 75
>3 754 >5
Figure 3-9. OC concentrations measured at CSN sites across the U.S., 2005-2007.
December 2008 3-47 DRAFT-DO NOT QUOTE OR CITE
-------
EC - Annual
0 0 O
EC{ugAn3): <=0.5 >0.5 to 1 >1 to 1.5
EC - Spring
99 ,31015
>I.U
EC - Fall
& Q 0
0
o
a
10
0 •
8„ o
a>6
o.° 0
V
o °
a • « a
* °e r
„ „°o 0
eC[ur»*ri3) <=0.5 >05B 1
>3
EC - Summer
"b
0 O
EC(u8ftw3} <30 5 >05 to I
* ' v' -i1
... - *• %¦/
(9 a OO Q
8 -OO • 0
* o ** a,¦
* " o e <3 as o
o 0 o o o Jr ® o
o © *«S °«°
dS ¦ ¦ .
>1.5-2
EC - Winter
^ o ®
O o
O 0
© b
* •
®0 ©
iCftigfliiSi <=06 j(t 6 w i
•. ^ f:«.s
* s j» 00 »
, 'i >4 ¦
• ° W
/HI ^ ii
©
>1 (01 5
>2
Figure 3-10. EC concentrations measured at CSN sites across the U.S., 2005-2007.
December 2008 3-48 DRAFT-DO NOT QUOTE OR CITE
-------
S04 - Annual
S04{ugftn3): <=2
>2 to 4
>4 to 6
804 - Spring
0 o
o O
°°o
o
go
%
0
<*>o
It
<8°°o"
°<£ o?qg
» °o8„°?
•5
a.
oo
o
>4 to 6
>6-8
a
• •
o
<3to
904 - Summer
w
; ° •
°e
o 0 °
a o
O
>4106
>6-8
SC»[ug(^3): <«2
>4»6
>2104
>2 to 4
Figure 3-11. SO42- concentrations measured at CSN sites across the U.S., 2005-2007.
December 2008 3-49 DRAFT-DO NOT QUOTE OR CITE
-------
N03 - Annual
8 Q
o
f
a ° w jt i 5 \ i «t
O /v. „ fc0. n o
8 _ CbO °0 T
0
C CP I1
<8®<
T
O O (p (O o
°e o 0 o
¦ *°s *«°
<0 0 •
.•» •
\ w
o o O
STN N03(ugfm3): <=1 >1to2 >2 to 3
N03 - Spring
cf
0°
©"
:* a
oO
V
cn0 uO
0 &8° _
® e °
° 0 o.n 45 !P ®
0 *" 0 o "If »
. i •»; ¦ >
o •
>1 toZ >?» 3 >3-4
N03 - Fal
•••
V
* ff..
3 0 (P 183 >?B 3 ->3-4 >4
STN WO^U9»m3) <®1
O
\
N03 - Winter
riW
V
• °
8
O
STM W03lygtYn3): <-1
Figure 3-12. NO3- concentrations measured at CSN sites across the U.S., 2005-2007.
December 2008 3-50 DRAFT-DO NOT QUOTE OR CITE
-------
NH4 - Annual
% tp
b
0 o
b 1 to 2
O #
>2 to 3 >3-4
NH4 - Spring
a.
° °
' . *p" <\'5*
„ t * 2to3
¦3J
NH4 - Fall
o
®o
b
0
&•
* ' Co%00 0
'¦8
0 •
>2to 3 >$4
- Summer
DO
%
B C
Qa
IM2 »2tt8
NH4 - Winter
. On tbti
\
I
o
' 9 *?" :0 °
°. * o„0 V p,
o * 'I# •
° D o. 0 ° 0
3 0 •
*f to 2 »2W3 >3-S
Figure 3-13. NH4+ concentrations measured at CSN sites across the U.S., 2005-2007.
December 2008
3-51
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Figure 3-9 shows regions of high OC with annual average concentrations greater than 5 (ig/m3 in
the western and the southeastern U.S. Concentrations at the western monitors peak in the fall and winter
while those in the Southeast peak anywhere from spring through fall. The central and northeastern
portions of the U.S. generally contain lower measured OC. Bell et al. (2007a) present a similar map for
estimated OCM (2000-2005) calculated by multiplying the blank corrected OC measurement by 1.4 to
account for non-carbon mass. This differs from the raw OC values that have not been multiplied by a
scaling factor, shown in Figure 3-9. There are a range of estimates in the literature for suggested scaling
factors (Turpin and Lim, 2001), depending predominantly on how highly oxygenated the aerosol is (Pang
et al., 2006). Fresh PM, more common in urban regions, has undergone limited chemical transformation.
As the aerosol is transported to rural regions, it becomes more oxygenated and hence heavier. As a result,
the necessary correction factor to account for non-carbon mass is higher in rural locations with estimates
ranging from 1.6 to 2.6 for IMPROVE monitors (El-Zanan et al., 2005). Therefore, applying one
correction factor of 1.4 across the entire U.S. will lead to an underestimate of the OCM in rural regions.
Therefore, the data presented in Figure 3-9 represent OC as measured and with a national blank
correction, but have not been adjusted to OCM by use of a scaling factor.
Figure 3-10 contains a similar map for EC that exhibits smaller seasonal variability than OC,
particularly in the eastern half of the U.S. There are isolated monitors spread throughout the country that
measure high annual average EC levels. These EC 'hot spots' are primarily associated with larger
metropolitan areas such as Los Angeles, Pittsburgh, and New York, but El Paso, TX, also reported high
annual average EC levels (driven by a wintertime average concentration greater than 2 jig/nr'). In a
similar analysis for EC by Bell et al. (2007a) for 2000-2005 data, there were also high wintertime EC
levels in eastern Kentucky and western Montana. These particular locations do not stand out in the
2005-2007 data in Figure 3-10.
Figure 3-11 contains a map for S042 which shows that S042 is more prevalent in the eastern U.S.
owing to the strong west-to-east gradient in S02 emissions. This gradient is magnified in the summer
months when more sunlight is available for photochemical formation of S042 . In contrast, N03 in Figure
3-12 is highest in the west, particularly in California. There are also elevated levels of N03 in the upper
midwest. The seasonal plots show generally higher N03 in the wintertime as a result of temperature
driven partitioning. Exceptions exist in Los Angeles and Riverside where high N03 readings appear year
round. The NH4 concentration maps in Figure 3-13 shows spatial patterns related to both S042 and N03
resulting from its presence in both (NH4)2S04 and NH4NO3. Annex A contains similar concentration maps
for Cu, Fe, Ni, Pb, Se and V as measured by XRF. There is considerably less seasonal variation in the
concentration profile for these metals than OC or the ions.
December 2008
3-52
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
For the fifteen metropolitan areas identified earlier, the contribution of the major component
classes to total PM2 5 mass was derived using the measured S042 , adjusted N03 . derived water, inferred
carbonaceous mass approach (SANDWICH) (Frank, 2006). This approach uses the measured FRM PM2 5
mass and co-located CSN chemical constituents to perform a mass balance-based estimation of the PM2.5
fraction attributed to S042 . N03 . EC, OCM, and crustal material. S042 and N03 include associated
NH4+ mass and estimated particle-bound water. Furthermore, N03 is assumed to be fully neutralized as
NH4NO3 and has been adjusted to represent the amount retained by the FRM monitor. EC is taken as
measured, and the crustal component is derived from common oxides contained in the Earth's crust
(Pettijohn, 1957). Finally, OCM is estimated using mass balance by subtracting the sum of all other
constituents from the FRM PM2 5 mass. The SANDWICH method takes into account passive collection of
semi-volatile or handling-related mass on the FRM filters in the mass balance calculation. The magnitude
of this artifact is assigned a nominal value of 0.5 (ig/m3 which is derived from limited analysis of FRM
field blanks. Other constituents such as salt and other metallic oxides, however, are not included in these
calculations and therefore the OCM fraction represents an upper bound on the FRM retained OC. The
calculations and assumptions that go into the SANDWICH method are discussed in detail in Frank (2006)
with further information available on EPA's AirExplorer web site
(http://www.epa.gov/cgi-bin/htmSQL/mxplorer/querv spe.hsql).
December 2008
3-53
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
v
] Sulfate
Nitrate
EC
Crustal
OCM
SANDWICH sulfate and nitrate include ammonium and water
Figure 3-14. Annual average FRM PM2.5 speciation data for 2005-2007 derived using the
SANDWICH method in fifteen CSAs/CBSAs: Atlanta, Birmingham, Boston, Chicago,
Denver, Detroit, Houston, Los Angeles, New York, Philadelphia, Phoenix, Pittsburgh,
Riverside, Seattle and St. Louis. Pie diameters are scaled to PM2.5 mass (ng/m3).
Figure 3-14 shows the PM2.5 compositional breakdown for the fifteen CSAs/CBSAs. All available
monitoring sites with co-located FRM PM2 5 and speciation data reporting in all four seasons for at least
one calendar year from 2005-2007 were included. Furthermore, each season was required to contain five
reported values for mass and the major PM2 5 constituents. This resulted in a varying number of sites
(ranging from one to seven) used to create the averages shown in Figure 3-14.
On an annual average basis, S042 is a dominant PM component in the eastern U.S. cities. For the
presented cities, this spans Houston to Boston where S042 makes up between 42 and 56% of PM25 on an
annual average basis. OCM is the next largest component in the east. In the west, OCM is the largest
constituent on an annual basis, ranging from 34% in Los Angeles to 58% in Seattle. S042 , NO, and
crustal material are also important in many of the included western cities. Fractional S042 ranges from
18% in Denver to 32% in Los Angeles. Fractional NO;, is relatively large in Denver (15%), Los Angeles
(19%) and Riverside (22%) and less important on an annual basis in Phoenix (1%) and Seattle (2%).
Crustal material is particularly prevalent in Phoenix (28%). EC makes up a smaller fraction of the PM2 5
(4 to 11%), but it is consistently present in all included cities regardless of region.
The seasonal variation in PM2 5 composition across the fifteen CSAs/CBSAs is shown in Figures
3-15 through 3-18 where the seasons are defined as before. S042 dominates in most metropolitan areas in
December 2008
3-54
DRAFT-DO NOT QUOTE OR CITE
-------
1 the summertime, while NOi becomes important in the colder wintertime months. Notable summertime
2 exceptions include Denver, Phoenix, and Seattle, where SOf makes up a smaller fraction of the PVI35
3 mass. Likewise, NO,' is less pronounced in the wintertime in Atlanta, Birmingham, Houston, Phoenix,
4 and Seattle. Los Angeles and Riverside exhibit elevated N03 from spring through fall. Crustal material is
5 a substantial summertime component in Houston (26%), and is generally low elsewhere in the East in all
6 seasons. In the West, crustal material represents a substantial component year-round in Phoenix and
7 Denver.
Winter
Sulfate
Nitrate
EC
OCM
Crustal
SANDWICH sulfate and nitrate include ammonium and water
Figure 3-15. Seasonally averaged FRM PM2.5 speciation data for 2005-2007 for winter derived using
the SANDWICH method in fifteen CSAs/CBSAs: Atlanta, Birmingham, Boston,
Chicago, Denver, Detroit, Houston, Los Angeles, New York, Philadelphia, Phoenix,
Pittsburgh, Riverside, Seattle and St. Louis. Pie diameters are scaled to PM2.5 mass
(ng/m3).
December 2008
3-55
DRAFT-DO NOT QUOTE OR CITE
-------
Spring
Sulfate
Nitrate
EC
OCM
Crustal
SANDWICH sulfate and nitrate include ammonium and water
Figure 3-16. Seasonally averaged FRM PM2.5 speciation data for 2005-2007 for spring derived using
the SANDWICH method in fifteen CSAs/CBSAs: Atlanta, Birmingham, Boston,
Chicago, Denver, Detroit, Houston, Los Angeles, New York, Philadelphia, Phoenix,
Pittsburgh, Riverside, Seattle and St. Louis. Pie diameters are scaled to PM25 mass
(ng/m3).
Summer
Sulfate
Nitrate
EC
OCM
Crustal
SANDWICH sulfate and nitrate include ammonium and water
Figure 3-17. Seasonally averaged FRM PM2.5 speciation data for 2005-2007 for summer derived
using the SANDWICH method in fifteen CSAs/CBSAs: Atlanta, Birmingham, Boston,
Chicago, Denver, Detroit, Houston, Los Angeles, New York, Philadelphia, Phoenix,
Pittsburgh, Riverside, Seattle and St. Louis. Pie diameters are scaled to PM2.5 mass
(jig/m3).
December 2008
3-56
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Fall
Sulfate
Nitrate
EC
OCM
Crustal
SANDWICH sulfate and nitrate include ammonium and water
Figure 3-18. Seasonally averaged FRM PM2.5 speciation data for 2005-2007 for fall derived using
the SANDWICH method in fifteen CSAs/CBSAs: Atlanta, Birmingham, Boston,
Chicago, Denver, Detroit, Houston, Los Angeles, New York, Philadelphia, Phoenix,
Pittsburgh, Riverside, Seattle and St. Louis. Pie diameters are scaled to PM2.5 mass
(ng/m3).
As noted in the 2004 PM AQCD (U.S. EPA, 2004), primary biological particles (PBPs), which
include microorganisms, fragments of living things, and organic compounds of miscellaneous origin in
surface deposits on filters, are not distinguishable in analyses of total OC. A clear distinction should be
made between PBP and primary OC produced by organisms (e.g. waxes coating the surfaces of
organisms) and precursors to secondary OC such as isoprene and terpenes. Indeed, the fields of view of
many photomicrographs of PM samples obtained by scanning electron microscopy are often dominated
by large numbers of pollen spores, plant and insect fragments, and microorganisms. Bioaerosols such as
pollen, fungal spores, and most bacteria are expected to be found mainly in the coarse size fraction.
However, allergens from pollens can also be found in respirable particles (Monn, 2001; Taylor et al.,
2002). Matthias-Maser et al. (2000) summarized information about the size distribution of PBP in and
around Mainz, Germany in what is perhaps the most complete study of this sort. Matthias-Maser found
that PBP constituted up to 30% of total particle number and volume in the approximate size range from
0.35 (.un to 50 |_im on an annual basis. Additionally, whereas the contribution of PBP to the total aerosol
volume did not change appreciably with season, the contribution of PBP to total particle number ranged
from about 10% in December and March to about 25% in June and October.
December 2008
3-57
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Ultrafine Particles
Very little is known about the spatiotemporal distribution or composition of ultrafine particles on a
regional scale. In an urban setting, a large percentage of ultrafine particles come from combustion-related
emissions from mobile sources (Sioutas et al., 2005). Ultrafine particle number concentrations drop off
quickly with distance from the roadway (Levy et al., 2003; Reponen et al., 2003; Zhu et al., 2005), and
therefore concentrations can be highly heterogeneous. Studies characterizing spatial variability in
ultrafines are currently limited to a handful of close proximity locations and therefore are discussed in the
next section on urban-scale variability. As for composition, OC makes up the majority of ultrafine
particles in most regions. Herner et al. (2005) reported a gradual increase in OC mass fraction as particle
size decreases from 1 |_im (20% OC) to 100 nm (80% OC) in the San Joaquin Valley of California. Sardar
et al. (2005) found OC to be the major component of ultrafine particles at four locations in California,
with higher OC mass fraction in the wintertime relative to summertime. EC and S042 were also present
in the ultrafine samples but at much smaller mass fractions; EC was present year-round whereas S042
had a summertime preference. More detailed chemical characterization of the OC fraction of ambient
ultrafine particles is extremely limited, but recent studies have identified specific organic molecular
markers affiliated with motor vehicle emissions including hopanes and polycyclic aromatic hydrocarbons
(Fine et al., 2004; Ning et al., 2007; Phuleria et al., 2007).
3.5.1.2. Urban-Scale Variability
PMw
PMio mass concentration has been shown to vary as much as a factor of five over urban-scale
distances of 100 km or less, and by a factor of 2 or more on scales as small as 30 km in an analysis of
California air quality (Alexis et al., 2001). This can be attributed to the rapid settling velocity and
resulting short atmospheric lifetime of the coarse-mode particles making up the majority of PMi0 mass.
As a result, local emission sources often dominate PM10 annual average mass at certain monitors. Data
from the fifteen CSAs/CBSAs were used to investigate urban variability in PM reported to the AQS
database. Maps of PM10 monitor locations and box plots of seasonal PM10 mass concentration data are
provided for Boston (Figures 3-19 and 3-20), Pittsburgh (Figures 3-21 and 3-22), and Los Angeles
(Figures 3-23 and 3-24). Annex A shows the PMi0 monitor locations and box plots of seasonal PMi0 mass
concentration for all 15 CSAs/CBSAs. Boston is an example of a city with a wide range in concentrations
measured at different sites. Inter-monitor variation in PMi0 is frequently larger than the seasonal variation
measured at any given site. Pittsburgh is an example of a city with a large number of PMi0 monitors
providing consistent values with a select few reporting higher concentrations (sites D, H, I and K in
December 2008
3-58
DRAFT-DO NOT QUOTE OR CITE
-------
Figure 3-22). This illustrates the potential influence of localized point or area sources or topography. Los
Angeles shows a high degree of between-season and within-season variability, which is on the order of
the between-monitor variation.
q) ~ Boston PM 10 Monitors
Boston Major Highways
| | Boston
0 10 20 40 60 80
i Kilometers
Figure 3-19. Map of PM10 FRM distribution with AQS Site IDs for Boston. MA.
December 2008
3-59
DRAFT-DO NOT QUOTE OR CITE
-------
AQS Site fD
Site A 254)234XM2
SteB 25-027-0023
SteC 33-011-0020
SifcB 334)154)014
SteE 4443034)002
SteF 44-007-0022
SteG 44*0074)026
SiteH 44-0074)027
1 -winter
2=spring
3=$umrm
4=fall
Figure 3-20. Box plot illustrating the seasonal distribution of 24-h average PM10 concentrations for
Boston, MA.
Table 3-14. Inter-sampler correlation statistics for each pair of PM10 AQS data for Boston, MA.
Site
A
B
C
D
E
F
G
H
A
1.00
0.69
0.69
0.73
0.71
0.84
0.70
0.79
(0.0,0.00)
(15.0,0.22)
(12.0,0.20)
(10.0,0.22)
(13.0,0.30)
(8.0,0.14)
(15.0,0.20)
(10.0,0.17)
191
169
179
173
171
182
169
167
B
1.00
0.66
0.56
0.45
0.69
0.77
0.65
(0.0,0.00)
(17.0,0.24)
(19.0,0.28)
(24.0,0.39)
(15.0,0.21)
(12.0,0.17)
(16.0,0.20)
174
167
161
158
169
156
154
C
1.00
0.72
0.47
0.62
0.64
0.59
(0.0,0.00)
(10.0,0.22)
(17.0,0.33)
(12.0,0.21)
(16.0,0.26)
(16.0,0.24)
182
170
168
179
166
164
D
1.00
0.63
0.68
0.59
0.69
(0.0,0.00)
(11.0,0.29)
(10.0,0.23)
(19.0,0.30)
(13.0,0.26)
175
163
173
161
158
E
KEY
1.00
0.84
0.58
0.80
Pearson R
(0.0,0.00)
(13.0,0.29)
(22.0,0.38)
(15.0,0.33)
(90th Percentile, COD)
171
171
161
157
F
n
1.00
0.81
0.95
(0.0,0.00)
(11.0,0.16)
(5.0,0.11)
182 169 167
G
1.00
0.79
(0.0,0.00)
(10.0,0.13)
169 154
H
1.00
(0.0,0.00)
168
December 2008
3-60
DRAFT-DO NOT QUOTE OR CITE
-------
Pi ttsb u rg h PM 10 M o n ito rs
Pi ttsb u rg h M ajo r H i g h way s
Pittsburgh
i Kilometer
Figure 3-21. Map of PM10 FRM distribution with AQS Site IDs for Pittsburgh, PA.
December 2008
3-61
DRAFT-DO NOT QUOTE OR CITE
-------
A
8
C
D
E
F
G
H
I
Mean
21,2
18,2
21.8
27.7
23.4
19.6
18,4
29.7
2S.2
Obs
1077
1019
1083
1087
176
179
978
182
1022
SD
12.9
11.4
12.3
20.3
11.1
9.9
11.7
16.8
19,3
Si® A
SiteB
SitaC
SiteD
Site E
SiteF
SiteG
SlteH
AOS Site ID
<2-00! 0002
42-0O3«B1
42-00 VOOd
42-9)10Cb4
42-833-0092
42-033-0095
42-053-0116
42-003-1301
90
80
70
S""
E 60
c
o
50-
40-
30-
c
80 -
SiteN
42-125-0005
ZL
SileO
70 -
42-125-5001
o
SiteP
42-129-000?
!XS
60 -
SiteO
42 129-0008
50 -
u
c
o
40 -
u
30 -
20 -
1 -winter
2ispiing
10-
3 summer
4 fall
o -
20.1
177
36.5
1061
263
1051
26,4
1069
21.7
1092
19.7
167
27.3
178
103
26,7
16.3
15,0
12.4
11,0
12.0
21,1
1079
11.4
In
in
1234 1234 1234 1234 1234 1234 1234 1234
Figure 3-22. Box plots illustrating the seasonal distribution of 24-h average PM10 concentrations in
Pittsburgh, PA.
December 2008
3-62
DRAFT-DO NOT QUOTE OR CITE
-------
Table 3-15. Inter-sampler correlation statistics for each pair of PM10 AQS data for Pittsburgh, PA.
bite
A B
C
D
b
h
(i
H
I
A
i.uu u.yj
u.yj
u.tsu
u.yz
u.yy
u.yj
u./y
U.bb
(u.u.u.uu) y
yb4
C
I.UU
U.bl
u.y4
u.yj
u.y4
u.//
U.8 /
(U.U.U.UU)
(23.UU.2U)
lb.U.U.11)
(/.U.U.12)
(8.U U.13)
(21.U.U.22)
(19.U U.1/)
1083
10/b
1/3
1/b
tlbb
1/9
101U
D
1.UU
U./2
U.bb
U./b
U.83
U.88
(U.U.U.UU)
(21 .U, U.2U)
(2b.U, U.24)
(2/.U, U.24)
(14.U, U.18)
db.U, U.U)
108/
1 /b
1/y
y/u
1&2
1014
t
I.UU
u.yu
u.yu
U./8
U.//
(U.U.U.UU)
(1U.U.U.14)
(1U.U.U.1/)
(2U.U, U.2U)
(2U.U.U.19)
KEY
176
173
154
175
166
F
Pearson K
1.UU
U.94
0.70
0.74
(90th Percentile, COD)
(0.0, 0.00)
(7.0,0.12)
(25.0, 0.27)
(25.0, 0.22)
n
179
157
178
168
G
1.00
o./o
0.8/
(U.U.U.UU)
(22.U, U.28)
(2U.U, U.19)
tJ/8
1&U
91U
H
1.UU
U./b
(U.U, U.UU)
(1/.U, U.2U)
182
1/1
I 1.UU
(U.U.U.UU)
1O22
Site
J
K
L
M
N
0
P
Q
A
U.84
U./b
U.88
U.8b
U.8b
u.//
U. / 8
U.8b
(14.0, 0.20)
(4U.U, U.3U)
db.U, U.18)
db.U, U.19)
(11 .U, U. 1 b)
db.U, U.22)
db.U, U.19)
(11.U, U.1b)
1/b
1U44
1U33
1Ub2
1U/4
1 bb
1//
1Ub1
B
U.93
U./b
U.bb
U.81
U.91
U./b
U.b3
U.bb
(/¦U, U.1 b)
(43.U U.3b)
(19.0, 0.23)
(2U.U, U.2b)
(1U.U, U.ib)
(12.U.U.19)
(18.0,0.28)
(10.0,0.18)
1b4
9iJb
9to
9/
ite
1003
C
U.9U
U./b
U.bb
U.83
U.89
U./8
U.88
u.yu
(8.U.U.13)
(39.UU.3U)
(14.U U.1/)
db.U, U.19)
(9.U.U.12)
(12.U.U.18)
(13.U, U.19)
(9.0,0.12)
1/4
1049
1039
10b/
108O
1&4
1/b
10b/
D
U./3
U.84
0.80
0./8
U./b
U.b/
U.b4
0./4
(24.U, U.22)
(24 U, U.22)
(2U.U, U.18)
(2U.U, U.2U)
(2b.U, U.2U)
(28.U, U.2b)
(2U.U, U.2b)
(2b.U, U.21)
1//
10bb
1U43
1Ub1
1U84
1b/
1/8
1U/1
E
U.8b
U.bb
U.83
U.8U
U.84
U.//
U.84
U.8b
(10.0, 0.1b)
(3b.U, U.29)
db.U, 0.1b)
(14.U, U.1 /)
(12.U, U.14)
(U.U, U.19)
(13.U, U.1b)
(11.U, U.1b)
1/1
1b9
1b9
1/2
1/b
1b1
1/2
1/4
F
0.90
U.b/
U.82
U./b
U.8b
U.83
U.84
U.8b
(/¦0, 0.12)
(41.0, 0.34)
(2U.U, U.2U)
(19.U, 0.22)
(11 .U, U.14)
(9.U, U.1b)
db.U, U.22)
(9.U.U.14)
1/4
1/2
1/2
1/b
1/9
1b4
1/b
1//
G
U.92
U./3
U.8/
U./8
U.89
U.81
U.84
U.8b
(/¦U, U.13)
(4b.U U.3b)
(18.U.U.21)
(19.U.U.24)
(9.U, U.ib)
(11 .U, U.1/)
(1/ U, U.2b)
(1U.U, U.ib)
1 bb
9bb
9!j8
9b2
y/b
14b
1b/
9(J/
H
U./4
U.b8
u.//
U. / 8
U./4
U.bU
U.bb
U./b
(23.U, U.26)
(2b.U, U.22)
db.U, U.18)
(1/.U.U.18)
(21 .U, U.22)
(2/.U.U.29)
(19.U, U.22)
(21 .b, U.24)
1/b
1/b
1/b
1/8
182
1b/
1//
18U
I
U./9
U.83
U.82
U. / 8
U.81
U.bb
U.b9
U./8
(22.U.U.2U)
(3U.U, U.2b)
db.U, U.1/)
(18.U, 0.20)
(20.U, U.1 /)
(2b.U, U.24)
(21.U, U.2b)
(22.U, U.19)
1 bb
992
9/8
998
1U19
1b8
1b/
1UU9
J
1.UU
U.bb
U./9
U./2
U.88
U. / 8
U.8b
U.8b
(U.U, U.UU)
(44.b, U.33)
(18.0, 0.20)
(18.U, 0.22)
(8.U.U.13)
(11 -U, U.1 /)
db.U, U.21)
(8.U, U.ib)
1//
1/U
1/0
1/3
1//
ito
1/3
1/b
K
1.UU
U./4
U./b
u./u
U. 4/
U.b8
U.b8
(U.U.U.UU)
(31.U U.2b)
(33.U U.24)
(4U.U, U.3U)
(44.U U.3b)
(34.U, U.3U)
(43.UU.3U)
l0b1
101 /
103b
l0b8
ito
1/1
1048
L
1.UU
U.8/
U.8b
U./U
U./4
U.8U
(U.U.U.UU)
(13.U, U.1 b)
db.U, U.1/)
(22.U, U.24)
(1/.U.U.21)
(18.U, U.19)
1Ub1
102b
1U48
1 bU
1/1
1U3b
M
1.UU
U./4
U.b4
U.b/
U.//
(U.U, U.UU)
(18.U, U.21)
(19.U, U.2b)
(1/ U, U.22)
(18.U, U.19)
1Ub9
1Ub/
1b3
1/4
1Ub3
N
1.UU
U./2
U.8b
U.8b
(U.U.U.UU)
(13.U.U.18)
(14.U.U.2U)
(1U.U.U.14)
1092
l(j/
1/8
10/b
O
1.UU
U./b
U.b9
(U.U, U.UU)
(18.U, U.2b)
(14.0 U.19)
1b/
1 to
ite
P
1.UU
U.84
(U.U.U.UU)
db.U, U.21)
1/8
1/b
Q
1.UU
(U.U, U.UU)
1U/9
December 2008
3-63
DRAFT-DO NOT QUOTE OR CITE
-------
n,
\
J
•-^4
~ LosAngelesPM 10 Monitors
Los Angeles M ajor Highways
LosAngeles
0 10 20
40
60
80
i Kilometers
Figure 3-23. Map of PM10 FRM distribution with AQS Site IDs for Los Angeles, CA.
December 2008
3-64
DRAFT-DO NOT QUOTE OR CITE
-------
Site A
SteB
SiteC
SiteD
SiteE
SiteF
SiteG
Mean
Obs
SD
AOS Site ID
06-037-0002
06-037-1103
06-037-4002
06-037 6012
06-037-9033
06-059-0007
06-059-2022
90 '
80 ¦
70 '
S~
E
60 ¦
c
O
1=winter
2=spring
3=summer
4=fall
50 ¦
40 -
30 *
20 ¦
10-
A
35.3
169
19.8
B
31.1
175
13.3
C
31.5
178
19.6
D
27.3
176
1&1
E
23.7
985
12.1
F
33.5
175
37.6
G
21.6
162
9.4
1234 1234 1234 1234 1234 1234 1234
Figure 3-24. Box plots illustrating the seasonal distribution of 24-h average PM10 concentrations
for Los Angeles, CA.
Table 3-16. Inter-sampler correlation statistics for each pair of PM10 AQS data for Los Angeles,
CA.
Site
A
B
C
D
E
F
G
A
1.00
073
0.44
0.73
0.47
0.41
0.65
(0.0,0.00)
(17.0,0.17)
(27.0,0.24)
(24.0,0.22)
(28.0,0.26)
(29.0,0.24)
(30.0,0.28)
169
153
154
157
169
155
143
B
1.00
0.61
0.57
0.52
0.42
0.73
(0.0,0.00)
(14.0,0.14)
(21.0,0.24)
(23.0,0.23)
(15.0,0.16)
(20.0,0.23)
175
159
159
173
162
149
C
1.00
0.65
0.43
0.93
0.73
KEY
Pearson R
(0.0,0.00)
(27.0,0.28)
(22.0,0.24)
(11.0,0.11)
(21.0,0.22)
178
158
176
159
148
D
(90th Percentile, COD)
n
1.00
0.70
0.65
0.57
(0.0,0.00)
(16.0,0.20)
(26.0,0.28)
(19.5,0.24)
176
175
161
150
E
1.00
0.29
0.38
(0.0,0.00)
(26.0,0.25)
(20.0,0.24)
985
173
159
F
1.00
0.65
(0.0,0.00)
(21.5,0.22)
175
150
G
1.00
December 2008
3-65
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Figures 3-25 through 3-27 illustrate the relationship between inter-sampler correlation and distance
between sites for PMi0 measurements obtained in Boston, Pittsburgh and Los Angeles. These three cities
are characterized by different topography, climatology, sources and composition of PM. Plots are
provided for all fifteen urban areas of interest in Annex A. In each plot, a large amount of scatter can be
observed. This is consistent with the seasonal box plots of concentration shown in Figures 3-20, 3-22, and
3-24. The Boston data exhibit the strongest relationship between inter-sampler correlation and distance,
with average inter-sampler correlation remaining higher than 80% when samplers are 44 km apart
(R2 = 0.61). The lowest correlations on this plot originate from comparisons between Site B (rural
Worcester, MA) and samplers located at Sites E (West Greenwich, RI) and G (Providence, RI). Boston is
subject to long range transport of sulfate, which is a regional pollutant and is a major component of PM2 5
and PM10 in the eastern U.S. The Pittsburgh data shows some lower inter-sampler correlations, with one
sampler pair having only 66% correlation within a distance of 2 km. On average, inter-sampler correlation
remained higher than 80% when samplers were also separated by 44 km, but in this case with much
greater scatter (R2 = 0.28) than observed in the Boston data. As seen for the Pittsburgh PMi0 box plots in
Figure 3-22, sites D, H, I, and K have elevated means and high variability that is driving the observed
scatter. These four sites are all located in mountainous suburbs of Pittsburgh (North Braddock, PA,
Liberty, PA, Lincoln Boro, PA, and Beaver Falls, PA, respectively), where emissions from steel
manufacturing and frequent stable conditions in the planetary boundary layer cause localized events of
elevated concentration. When those four sites are removed, scatter decreases greatly (R2 = 0.56). The Los
Angeles data exhibit a much steeper slope, with average inter-sampler correlation remaining higher than
80% when samplers are only 30 km apart (R2 = 0.56). The lower inter-sampler correlations in part reflect
the fact that some of these monitoring sites are separated from each other by hills or, in the case of one
sited at Lancaster, CA, by the San Gabriel Mountains. The Los Angeles data exhibit greater scatter than
the Pittsburgh data. However, the smallest inter-sampler separation distance is 23 km, and there are
relatively fewer PMi0 samplers. It is not possible to judge how data would correlate on smaller spatial
scales.
December 2008
3-66
DRAFT-DO NOT QUOTE OR CITE
-------
1
0.2
0 4 1 , 1 1 , 1 1 1 ,
0 10 20 30 40 50 60 70 80 90 100
Distance Between Samplers (km)
Figure 3-25. PM10 inter-sampler correlations as a function of distance between monitors for
Boston.
0.8
~ ~ ~ ~
~ ~ ~ •
~ f ~ ~~ ~ ~ A #
% ~ ~ **~~~~~*
. ~ A ~ ~ A ~ *
~ "
~
!~
~ ~
~
~ ~~
~ 4
~ ~
~
T
~
~
~
0.4 -
30 40 50 60 70
Distance Between Samplers (km)
Figure 3-26. PM10 inter-sampler correlations as a function of distance between monitors for
Pittsburgh.
December 2008
3-67
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
0.8
0.6
0.4 -
0.2
~
~
10
20
30
40
50
60
70
80
90
100
Distance Between Samplers (km)
Figure 3-27. PM10 inter-sampler correlations as a function of distance between monitors for Los
Angeles.
A table of complete pair wise, within-city comparisons for PMi0 measured at the available monitors
in each CSA/CBSA are included for Boston (Table 3-14), Pittsburgh (Table 3-15) and Los Angeles (Table
3-16); the complete set of tables for all 15 CSAs/CBSAs are shown in Annex A. Comparison statistics
include the Pearson correlation coefficient (R), the 90th percentile of the absolute difference in
concentrations (P90), the coefficient of divergence (COD) and the number of paired observations
(Wongphatarakul et al., 1998). The COD provides an indication of the variability across the monitoring
sites in each CSA/CBSA and is defined as follows:
COD]k=.
i j^( x„~xy
%
KX,+X'J
Equation 3-1
where Xu and Xik represent observed concentrations averaged over some measurement averaging period
(hourly, daily, etc.), for measurement period /' at sites j and k, and p is the number of paired observations.
A COD of 0 indicates there are no differences between concentrations at paired sites (spatial
homogeneity), while a COD approaching 1 indicates extreme spatial heterogeneity.
December 2008
3-68
DRAFT-DO NOT QUOTE OR CITE
-------
Boston
Boston
Boston
Kilometers
Pfv! 2.5 Monitors
Major Highways
Figure 3-28. PM2 5 monitor distribution in comparison with source distribution, Boston, MA.
December 2008
3-69
DRAFT-DO NOT QUOTE OR CITE
-------
SteA
SteB
Site C
SteO
SiteE
Stef
SiteG
SteH
Site i
Site J
A
8
C
D
e
F
G
H
J
Mean
9.t
31
8,3
94
9.6
11,7
n/>
30.S
121
107
Ote
Ml
W
342
355
357
349
m
349
tots
^35
sD
ExO
o.5
63
66
6wO
7-U
68
<3.9
69
7,2
AUS Stfv ID
?SCCb 1004
2S-W 2€0ft
25 Ctf> "CDS
25O2_,-00Q4
25 0^ COO?
25 0"5 0027
25 phxw;
25 orvoow
25-027-0016
4 0
XT
£
|
c
o
% 20 1
oj
u
c
o
ig-
1 =winter
2^-spiing
3-sumrner
4—fall
1234 1234 1234 1234
K
Mean 11.4
Ofc 346
SD 7.4
40
234 1234 1234 1234 1234 1234
L
7,2
183
53
M
10.0
361
6.7
N
ft?
362
6.0
0
8.9
362
5.8
P
10,1
1027
6.6
Q
11.9
32!
6.9
R
105
313
as
s
97
«e
6.5
SteK
Site L
Site M
SiteN
SiteO
SSbP
SiteQ
Site ft
Site S
AOS Site ID
3W01 2004
13 on 1015
.W-OU 1006
3301VQO14
44-0070022
44-007 0026
44-007 0028
44-0074010
SP
£
3
c
o
30-
20
Q)
U
C
o
u
10
}-winter
2=sprir»g
3=sumroer
4=fall
'I
1234 1234 1234 1234 1234 12 34 12 3 4 1234 12 3 4
Figure 3-29. Box plots illustrating the seasonal distribution of 24-h average PM2.5 concentrations
for Boston, MA.
December 2008
3-70
DRAFT-DO NOT QUOTE OR CITE
-------
Table 3-17. Inter-sampler correlation statistics for each pair of PM2.5 AQS data for Boston, MA.
bite
A B
C
U
t
h
Cj
H
I
J
A
too o.ao
o.//
071
0.84
(1/9
(1/8
(1/9
(1/9
(I.//
(0.0,0.00) (6.6,0.21)
(6.2,0.22)
(6.y, 0.2d)
(4.8,0.19)
(8.1,0.23)
(/./, 0.24)
(6.8,0.22)
(/. 9,0.25)
(/.5,0.24)
341 326
318
323
329
318
319
325
338
310
B
1.00
0.92
0.8/
0.8/
0.90
0.90
0.90
0.90
0.85
(0.0,0.00)
(4.1,0.10
(4.1,0.18)
(47,0.19)
(6.3,0.21)
(6.2,O.Zd)
(4.9,0.19)
(/.1,0.26)
(5.5,0.21)
350
328
331
if)
326
323
333
343
31/
C
1.00
uyj
u.os
u.yu
u.oy
u.yu
U.00
U.8b
(0.0,0.00)
(d.5, U.I/)
(5.d,0Z1)
(b.d,U.ZJ)
(b.d, U.^4)
(5.0.U.ZU)
(b.a.oA)
(6.2,0.21)
342
321
dd1
dlb
318
326
ddb
311
L)
1.00
m
0.88
0.88
0.8b
U.Ob
0.8/
KEY
(0.0,0.00)
(5.6,0.20)
(5.8,0.21)
(5.8,0.22)
(4.6,0.19)
(7.0,0.26)
(5.8,0.19)
hfearsonK
355
336
324
329
332
345
313
E
(90lh Percentile, CXDD)
1.00
0.90
0.90
0.89
0.87
0.87
(0.0,0.00)
(5.9,0.19)
(5.8,0.21)
(5.0,0.19)
(6.9,0.24)
(5.4,0.20)
357
330
333
340
350
322
F
1.00
0.94
0.94
0.92
0.92
(0.0,0.00)
(d.8,0.14)
(d.5,0.15)
(4.5,0.1/)
(5.4,0.18)
340
324
324
it)
310
G
1.UU
0.94
0.94
0.89
(0.0,0.00)
(4.0,0.16)
(4.d,0.15)
(57,0.20)
398
325
338
308
H
1.00
u.yj
u.oy
(0.0,0.00)
(4Y, U.iy)
(5.0,0.1/)
349
342
318
I
1.00
0.8b
(0.0,0.00)
(b.y,u.zd)
1015
ddO
J 1.00
(0.0,0.00)
3&
bite
K
L
M
N
0
V
Q
K
S
A
07/
0.b1
071
0.68
O./d
0.8/
0.81
0.85
0.86
(8.1,0.23)
(8.3,0.29)
(8.0,0.23)
(/.90.23)
(/.0,0.22)
(540.18)
(/.20.23)
(5.6,0.20)
(5.2,0.18)
320
1/3
324
334
331
326
292
285
306
B
0.8b
0.80
0.8/
0.83
0.88
0.86
0.80
U.Ob
0.85
(b.b,0.21)
(6.2,0.2d)
(6.0,0.21)
(4./, 0.18)
(5.6,0.19)
(/.y,0.26)
(57,0.21)
(6.0,0.19)
329
1/5
331
341
336
335
300
288
314
C
U.Ob
0.89
0.93
0.90
0.93
0.83
0./9
0.81
0.82
(b.y,uzn
(4.8,0.23)
(4.4,0.1/)
(4.b, u.iy)
(d.8,U.1S)
(s.y.u.zi)
(/.8, u.zto)
(b.z,uzd)
(6.0,0.21)
jtf
1/d
616
dd5
616
te)
290
ZU1
L)
u.oo
u./y
u.yi
U.85
U.8b
u.yu
U./t)
u./y
0.80
(b.4, u.iy)
(57, U.Z5)
(d.5, U.lb)
(47, u.iy)
(4/U.18)
(b.2, U.2U)
(/.8,u.a)
(6.2,0.21)
(5.8, U.A))
dZ3
1/4
32y
3dy
dd4
d4^
dUU
16!
d21
t
u.o/
0 .(Z
u.od
u./y
U.84
u.yi
U.Ob
0.88
u.yi
(64,0.20)
(8.3,0.2/)
(5.8,0.1/)
(6d, U.2U)
(4.8U.18)
(4.b, U.1/)
(6.d, 0.22)
(4.9,0.18)
(3.9,0.1/)
333
1/9
338
34/
343
343
306
295
324
F
0.91
0/8
0.90
U.85
U.85
u.oy
U.bb
U.00
u.oy
(47,0.1/)
(y.6,o.aj)
(54 0.18)
(6.4, U.2U)
(/.5,022)
(5.2,U.16)
(6.0,0.16)
(4.9,0.16)
(b.b, U.1/)
323
168
323
334
330
336
&
281
316
G
0.90
07/
0.90
U.Ob
0.8/
U.OO
U.bb
0.8/
U.00
(5.u, u.iy)
(y.u, u.jj)
(b.d, U.2U)
(/.0,0.22)
(5.5, U.I/)
(5.d,U.I/)
(5/U.1/)
(D./,u.iy)
d A)
VI
d
-------
Philadelphia PM 2.5 Monitors
Philadelphia Major Highways
Philadelphia
d)
0 10 20
40
60
80
i Kilometers
Figure 3-30. PM2 5 monitor distribution in comparison with source distribution, Pittsburgh, PA.
December 2008
3-72
DRAFT-DO NOT QUOTE OR CITE
-------
Mean 15,1
Obs 1063
SD 8,9
Site A
SlteB
S'tmC
SteO
Site'E
Site F
SteG
SiteH
Site I
Site J
SiteK
Site L
AQS Site ID
42-001 J007
42 «07 0014
42I2SOOOS
42-125 0200
42 125 4001
42-129-0006
1-winter
2-spring
3-iurnmer
4 .-fall
B
19.8
1066
14.7
13,2
13.6
15.1
306
8.0
165
as
332
9,0
16.4
337
9.5
AGS Site ID
42-003-0006
42*003-0064
42-00XX»7
42-003-0095 «r»
42-003-1006 Jj.
42 003 1301 jl
c
,2
'4Z
to
c
u
c
o
u
1=wi titer
2=spriog
3=summer
4=fail
70
60 "
SO
40 "
30
20 -
1 0
12 3 4
12 3 4
12 3 4
12 3 4
12 3 4
12 3 4
G
H
I
J
K
L
Mean 15.3
16.4
15,5
14.8
13.4
15.4
Obs
171
328
354
345
966
350
SD
8,3
9.3
8.6
ao
8.6
a 7
sr
E
c
o
c
at
kj
c
o
u
70
6 0
SO '
40 •
30
20
1 0
ll'l
1234 1234 1234 1234 1234 1234
Figure 3-31. Box plots illustrating the seasonal distribution of 24-h average PM2.5 concentrations
for Pittsburgh, PA.
December 2008
3-73
DRAFT-DO NOT QUOTE OR CITE
-------
Table 3-18. Inter-sampler correlation statistics for each pair of PM2.5 AQS data for Pittsburgh, PA.
A
B C
D
E
F
G
H
I
J
K
L
A
1.00
o./y o.yb
0.92
0.93
0.9b
0.9b
0.8b
o.yo
0.93
0.91
0.88
(0.0,0.00)
(16.9,0.19) (b.6, 0.13)
(4./, 0.11)
(47,0.11)
(4.9,0.10)
(3.8,0.10)
(6.4,0.13)
(6.4,0.13)
(b.0,0.12)
(6.0,0.13)
(b.6,0.12)
106J
itiib i9H
164
'515
&y
1/0
319
344
'661
934
340
B
1.00 0./1
0.6b
0.80
U.bb
0./6
0.69
0/1
U.bb
0.68
0.6/
(0.0,0.00) (16.9,0.24)
(1/.4,0.2b)
(14.4,0.19)
(12.b, 0.14)
(1b./,0.20)
(1/0,0.19)
(1b./, 0.21)
(1/.80.23)
(19.3,0.2b)
(1b.9,0.21)
ite 3to
1&
329
3&
1/1
3^4
3b0
341
9^8
346
u
1.00
O.bO
0.90
u.yi
0.94
O.iD
U.bW
0.96
0.9b
0.91
(0.0,0.00)
(2.8,0.09)
(b.b, 0.1b)
(B./,0.1/)
(6.0,0.14)
(9.4,0.19)
(6./, 0.1b)
(4.6,0.12)
(4b, 0.10)
(6.b, 0.1b)
AJ6
-(44
^82
^82
148
m
'Ml
'1m
ita
^86
L>
1.00
0.84
0.8/
0.91
0/9
0.89
0.91
0.9/
0.8b
(0.0,0.00)
(b.4,0.1b)
(H.b.O.lb)
(b-8 0.1 J)
(9.2,0.1/)
(b-9,0.1 J)
(4.6,0.11)
P.1.0.0H)
(6.b, O.lb)
16b
1b3
1b1
1b8
1b6
1b8
1bb
146
1b/
b
1.00
0.90
0.90
0.84
0.8b
0.86
0.88
0.83
(0.0,0.00)
(6.4,0.13)
(6.b, 0.13)
(6.8,0.14)
(8.3,0.16)
(/./, 0.16)
(/.6,0.1b)
(1.5,0.1b)
KEY
332
313
157
29b
320
31b
290
318
h
htearsonK
1.00
0.91
0.82
0.88
0.88
0.89
0.86
(90th Rsrcentik, CXDD)
(0.0,0.00)
(6./, 0.13)
(/.4,0.14)
(/.1,0.1b)
(/.9,0.1b)
(8.8,0.1/)
(/-0,0-14)
16/
302
'511
319
M)
m
G
1.00
0.78
0.94
0.93
0.90
0.91
(0.0,0.00)
(/.3.0.16)
(4.0,0.10)
(b.0,0.11)
(6.6,0.1b)
(b.0,0.13)
1/1
1b9
16J
1b9
149
161
H
1.00
0.80
0./8
0.82
o./o
(0.0,0.00)
(8.4,0.1b)
(8.2,0.1/)
(9.0,0.18)
(9.2,0.18)
^28
'Ml
^09
^88
^14
I
1.00
U.bW
u.by
U.bb
(0.0,0.00)
(b.0,0.11)
(/20.16)
(6.0 0.13)
&4
&4
^10
to9
J
1.00
u.ya
0.88
(0.0,0.00)
(b.b, D.TZ)
(b-9,0.15)
34b
302
331
K
1.00
0.86
(0.0,0.00) (6.9,0.1b)
966
306
L 1.0T
December 2008
3-74
DRAFT-DO NOT QUOTE OR CITE
-------
rJ
LosAngeles PM 2.5 Monitors
Los Angeles Major Highways
LosAngeles
0 10 20 40 60
80
i Kilometers
Figure 3-32. PM2.5 monitor distribution in comparison with source distribution, Los Angeles, CA.
December 2008
3-75
DRAFT-DO NOT QUOTE OR CITE
-------
Site A
Site 8
Site
Step
SteE
SteF
SiteG
SteH
Site t
Site J
SiteK
AQS Site ID
OMM7O0O2
0MB7 1002
0WS7-1KB
06-037-1201
06-037 1301
06037-2005
QMB7-4002
06037 «XK
06-037-9033
0M5MJ0Q?
06-059-2022
1 =winter
2=spring
3=summer
4=fall
—
£
Mean
Ob
su
60
SO'
40'
c
O
% 30'
*-
c
at
u
C
O 20'
u
10
A
16.1
862
to .a
B C
17.0 16.7
306 1004
10.2 9.8
0
13.3
291
7,5
£
W
32?
9.3
F
143
334
8.9
G H
14.7 14.2
946- 990
8.4 7.?
8.2
221
W
J
14.4
999
8.5
K
10.9
318
6.4
1234 1234 1234 1234 1234 1234 12 3 4 1234 1234 12 3 4 1234
Figure 3-33. Box plot illustrating the seasonal distribution of 24-h average PM2.5 concentrations for
Los Angeles, CA.
Table 3-19. Inter-sampler correlation statistics for each pair of PM2.5 AQS data for Los Angeles,
CA.
A
B
C
U
b
h
G
H
I
J
K
A
1.00
0.8b
0.8/
0.81
0.80
0.88
0.b8
0.b4
0.30
O./O
0.82
(0.0,0.00)
(9.0,0.18)
(/./,0.1b)
(9.0,0.19)
(9./, 0.21)
(5.B.U.14)
(11.b,0.Z4
(1Z4,U,U)
(18.0,0.3b)
(lO.b, 0.21)
(11.4,0.23)
8b2
2b2
803
238
2b2
2b9
/b1
/93
1/9
804
2b9
B
1.00
0.92
0.8/
0.83
0.88
0.//
0./3
0.31
0/4
0./1
(0.0,0.00)
(b.b, 0.11)
(9.1,0.19)
(9.0,0.1b)
(/.b, 0.1b)
(9.8,0.1/)
(11.b, 0.18)
(24.1,0.38)
(11.9,0.19)
(1b.0,0.2/)
308
293
2b0
2/8
2b8
282
1//
292
2//
(J
1.00
0.80
0.89
u.y^
0.84
0/9
0.29
0.82
0./8
(0.0,0.00)
(9.B, 0.20)
(b.8,0.11)
(6.4,0.13)
(9.0,0.1b)
(10.0,0.1/)
(18.b, 0.38)
(9.4,0.1b)
(13.2,0.2b)
1004
2/4
31b
319
880
913
213
920
30b
L)
1.00
0.b9
0.//
0.b3
O.bO
0.41
0.b4
O.bO
(0.0,0.00)
(10.9,0.23)
(/.4.0.18)
(11.3,0.22)
(11.1,0.22)
(14.8,0.31)
(9.6,0.21)
(11.60.23)
291
2b3
2b3
2bb
2b8
1b4
2/4
2b1
b
1.00
0/9
0.9b
0.92
0.34
0.88
O./b
(0.0,0.00)
(9.1,0.19)
(b.9,0.11)
(/.b, 0.13)
(19./, 0.39)
(8.2,0.1b)
(13./,0.2/)
32/
301
289
301
192
30/
291
h
1.00
O./O
o./o
0.33
0.b9
0./2
(0.0,0.00)
(10.b, 0.18)
(9.2,0.19)
(14.8,0.34)
(9.8,0.19)
(9.9,0.21)
334
290
302
184
311
293
U
1.00
0.9b
0.23
0.92
0/8
(0.0,0.00)
(4.0,0.09)
(1/.0,0.3b)
(b.4,0.12)
(11.0,0.21)
94b
8b9
194
882
2//
H
1.00
0.2b
0.91
0./8
(0.0,0.00)
(1b.3,0.34)
(b.9,0.12)
(9b, 021)
990
208
914
294
I
1.00
0.21
0.31
(0.0,0.00)
(18.3,0.3b)
(9./, 0.28)
221
20b
180
J
1.00
0.84
(0.0,0.00) (9.8,0.19)
999
298
nr
~31F
December 2008
3-76
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
PMzs
PM2 5 has a longer residence time in the atmosphere resulting from a slower settling velocity
compared to PMi0. As a result, increased spatial homogeneity is expected with less localized influence
from point sources. Maps of PM25 monitor locations and box plots of seasonal PM2.5 mass concentration
data are provided for Boston (Figures 3-28 and 3-29), Pittsburgh (Figures 3-30 and 3-31), and Los
Angeles (Figures 3-32 and 3-33) for direct comparison with the PMi0 maps and box plots provided in
Figures 3-19 through 3-24. Annex A shows the PM2 5 monitor locations and box plots of seasonal PM2 5
mass concentration for all 15 CSAs/CBSAs. With very few exceptions, the PM2 5 is quite uniformly
distributed across the monitors. Boston, Los Angeles, New York and Phoenix all have one monitor that is
recording visually less PM2 5 than the rest of the monitors within the respective CSAs/CBSAs; Riverside
has two such monitors. Unlike PMi0, PM2 5 varies approximately the same magnitude between monitors
as it does between seasons for the 15 selected cities.
Figures 3-34 through 3-36 show the relationship between inter-sampler correlation and distance for
PM2 5 measurements obtained in Boston, Pittsburgh, and Los Angeles to illustrate how this relationship
varies across urban areas with different topography, climatology, and sources in a way similar to that for
PM10 as shown in Figures 3-25 through 3-27. Plots are provided for the fifteen CSAs/CBSAs
characterized here in Annex A. In each plot, substantially less scatter is observed in these plots when
compared with those for PMi0. This is consistent with the seasonal box plots of concentration shown in
Figures 3-29, 3-31, and 3-33. The Boston data exhibit the strongest relationship between inter-sampler
correlation and distance, with average inter-sampler correlation remaining higher than 80% when
samplers are 95 km apart (R2 = 0.55). This small amount of variability is sensible given the consistency
between distributions shown in the box plots. The Pittsburgh data show some reductions in inter-sampler
correlations at short distances, with the samplers at Sites B (Liberty, PA, described above for PMi0) and G
(in the neighboring town of Clairton, PA located at a lower elevation on the bank of the opposite side of
the Monongahela River from Liberty) having only 76% correlation with a distance of less than 4 km. On
average, inter-sampler correlation remained higher than 80% when samplers were separated by 61 km,
but in this case with much greater scatter (R2 = 0.22) than observed in the Boston data. This scatter is
driven by the measurements at Site B; Figure 3-31 shows an elevated mean and variability compared with
other monitors situated around the Pittsburgh CSA. When data from Site B are removed, the inter-sampler
correlation vs. distance plot for Pittsburgh PM2 5 resembles the one from Boston (with R2 increasing to
0.68). The Los Angeles data exhibit a much steeper slope, with average inter-sampler correlation
remaining higher than 80% when samplers are 29 km apart (R2 = 0.74). This distance is similar to that for
December 2008
3-77
DRAFT-DO NOT QUOTE OR CITE
-------
1 the PM10 data, which suggests that other factors, such as mountainous topography separating monitors,
2 the distribution of traffic and suspension of crustal components, and occurrence of stable boundary layers,
3 may cause more spatial variation in the PM concentration profile within the Los Angeles region when
4 compared with other parts of the country. The Site I monitor at Lancaster CA, separated from the rest of
5 the Los Angeles region by the San Gabriel Mountains, again provides the low correlations concentrated in
6 the lower right portion of Figure 3-36. Tables 3-17 through 3-18 contain the pair-wise statistics (R, PSO,
7 COD and N) for PM2 5 in Boston, Pittsburgh, and Los Angeles, respectively. Annex A contains the
8 complete list of pair-wise statistics for PM2 5 measured within the fifteen CSAs/CBSAs.
1 f
~ ~
:
0.8
<
~ ~
~ ~
• . »> .. ..
^ ~ /~. ~
* A
0.6
o
o
0.4
0.2
0 4 1 1 1 1 1 1 1 1 1 1
0 10 20 30 40 50 60 70 80 90 100
Distance Between Samplers (km)
Figure 3-34. PM2.5 inter-sampler correlations as a function of distance between monitors for
Boston.
December 2008
3-78
DRAFT-DO NOT QUOTE OR CITE
-------
.4 --
.2 --
0 -
0
35,
1
.8 --
.6 --
.4 --
.2 --
0 -
0
10 20 30 40 50 60 70
Distance Between Samplers (km)
80 90
100
PM2.5 inter-sampler correlations as a function of distance between monitors for
Pittsburgh.
~
~ «~
* «»~ ~~ .
~ %
~
~
~
~
10 20 30 40 50 60 70
Distance Between Samplers (km)
80 90
100
PM2.5 inter-sampler correlations as a function of distance between monitors for Los
Angeles.
3-79
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
PMio-2.5
Given the limited number of co-located FRM PMi0 and FRM PM2 5 monitors, only a very limited
investigation into the intra-urban spatial variability of PM10-2.5 was possible. Of the 15 cities under
investigation, only six (Atlanta, Boston, Chicago, Denver, New York and Phoenix) contained co-located
FRM monitors adequate for generating PMi0_2 .5. Of these six cities, only Boston and New York had more
than one qualifying location within the CSAto allow for calculation of comparison statistics similar to
those above for PMi0 and PM2 5. Figure 3-37 contains box plots for all six CSAs/CBSAs and Annex A
contains the correlation tables for Boston and New York (two sites each). For Boston, the correlation
between the two sites for PM10-2.5 was 0.45 compared with 0.84 for PMi0 alone and 0.73 for PM2 5 alone
(using the same two monitoring sites). For New York, the correlation was slightly higher for the two sites:
0.74 for PM10-2.5 compared with 0.82 for PM10 alone and 0.93 for PM2 5 alone. The COD for PM10_2.5 also
increases in both cities compared with PM10 and PM2 5 alone, suggesting less spatial homogeneity. Some
of the disparity, however, could also be coming from the 2-monitor subtraction method used in calculating
PMio-2.5-
Atlanta
Boston
Chicago
Denver
New York
Phoenix
AQS Site P
13-121-0032
A
Me an 9 "
Obs! 167
SPj 6.6
30 "j
jrsy *
AOS SitelD
AQS Site ID
SieA
08-001-0005
A
18-127-0024 5
Mean
Mean
Moan
*•//*
///' ///'
AOS Sit* ID
'Mt" -1
04-013-9997
A
Mean
22.4
Obs
163
SD
11.6
8
-Hi
/
Figure 3-37. PM10.2.5 generated from all available co-located FRM PM10 and PM2.5 monitors in
Atlanta, Boston, Chicago, Denver, New York and Phoenix, 2005-2007. Box plots show
the median and interquartile range with whiskers extending to the 5th and 95th
percentiles at each site during (1) winter (December-February), (2) spring (March-May),
(3) summer (June-August) and (4) fall (September-November).
December 2008
3-80
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
PM Constituents
The pie charts showing PM composition that were generated using the SANDWICH method for
the fifteen CSAs/CBSAs presented earlier in Figures 3-14 through 3-18 represent a spatial average across
each area. Similar pie charts for each individual monitor within the fifteen CSAs/CBSAs are contained in
Annex A. In general, these charts reveal spatial homogeneity in PM25 bulk chemistry within each
metropolitan area. Some notable exceptions do exist, however. Birmingham and Detroit show variation in
the amount of crustal material, both spatially and seasonally. The Denver map reveals some spatial
variation in N03 during the winter, the season with the highest measured PM2 5 mass. Several sites in the
New York and one in Pittsburgh have elevated percentages of EC relative to the other sites within the
respective cities. The excess EC at the Pittsburgh site is associated with higher PM2 5 mass measured at
that site relative to others. In Phoenix, high winter PM2 5 mass is site specific and appears to be associated
with high organic carbon; the crustal component also varies and is inversely proportional to total
measured mass.
Ultrafine Particles
Only a handful of studies have performed direct comparisons of ultrafine measurements at multiple
locations within an urban center. An early study by Buzorius et al. (1999) suggested spatial homogeneity
in total particle number concentrations between multiple locations in Helsinki, Finland. They found
correlations in 10 min averages at three sites within the city as high as 0.84. The sites, however, were
relatively close together (2 km) and all near the same roadway. There was a high degree of correlation
between traffic intensity and total aerosol number concentrations, suggesting that traffic was the primary
source of the measured particles and the driving force behind the high correlations. Weekend correlations
(0.28-0.47) and correlations with a fourth monitor located 22 km outside the city (0.05-0.64) were much
lower.
Tuch et al. (2006) found more spatial heterogeneity in ultrafine particle concentrations measured
for an entire year at two locations 1.5 km apart in Leipzig, Germany. Figure 3-38 shows the correlation as
a function of particle size (mobility diameter) dropping off as the particle size decreases. Particles less
than 100 nm show correlations dropping from 0.5 down to 0.2. Annex A contains correlation coefficients
of hourly and daily average particle number, surface area and volume concentrations as a function of
particle diameter adapted from the Tuch et al. (2006) study. For all days (N = 5481 hourly observations),
the correlation between ultrafine particles (10-100 nm) measured at the two sites was 0.31.
December 2008
3-81
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
b io
c
O
07
o
O
c
o
TO
0
0
CJ
c
03
£
1
CO
-------
1
2
3
4
5
6
7
8
9
10
11
12
Influential parameters include canyon height to width ratio (H/W), source positioning, wind speed and
direction, building shape and upstream configuration of buildings. Figure 3-39 shows concentrations
obtained from wind tunnel and computational fluid dynamics simulations of transport and dispersion in
an infinitely long street canyon with a line source centered at the bottom of the canyon (Xiaomin et al.,
2006). When the canyon height was equal to the street width (typical of moderate density suburban or
urban fringe residential neighborhoods) and lower background wind speed existed, concentrations on the
leeward canyon wall were four times those of the windward wall near ground level. When the canyon
height was twice the street width (typical of higher-density urban planning) and background winds were
somewhat higher, near ground-level concentrations on the windward canyon wall were roughly three
times higher than those measured at the leeward wall. These results suggest that micro- and
neighborhood-scale variation related to urban topography may have a significant impact on airborne PM
exposures.
windward, B leeward,
measured measured
SSI!!!1 leeward,
si mutated simu|a(ed
0.8
0.8
0.6
¦'>
0.4
0.4
0.2
0.2
100
100 200 300 400 500 600
Dimensionless concentration
Dimensionless concentration
Source: Xiaomin et al. (2006)
Figure 3-39. Dimensionless concentration as a function of height at windward and leeward
locations and street canyon aspect ratios (H/W). (a) Dimensionless concentration on
the windward and leeward sides of the canyon when H/W = 1 and wind speed = 3 m/s.
(b) Dimensionless concentration on the windward and leeward sides of the canyon
when H/W = 2 and wind speed = 5 m/s. Computational fluid dynamics modeling was
performed, and measurements were obtained in wind tunnel simulations.
December 2008
3-83
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
The particle concentration profile is affected by land and building topography, meteorology,
particle size distribution, particle composition, and particle volatility. For instance, Viana et al. (2008)
observed higher concentrations of crustal elements in PM10 and PM2 5 samples in rural neighborhoods and
higher concentrations of combustion-derived PMi0 and PM2 5, such as EC and NO;, . in higher density
urban areas. Gutierrez-Daban et al. (2005) examined the mass distribution of various PAHs under
different traffic and urban density conditions. Figure 3-40 displays the distributions for benz[a]pyrene
(BaP) at high and low traffic sites at the urban center, periphery, and industrial areas found in
Gutierrez-Daban et al. (2005). It can be seen that concentrations were nearly an order of magnitude lower
for the low traffic urban periphery location when compared with the high traffic or industrial locations.
Particles smaller than ~ 600 nm had roughly an order of magnitude higher concentration than those at
larger sizes and tended to have a larger spread in concentrations among sampling sites. Figure 3-41 shows
the distributions for sixteen PAHs at a high traffic location at the city center from Gutierrez-Daban et al.
(2005). PAH species varied in concentration by up to two orders of magnitude for each particle size bin,
and the highest concentrations of individual PAHs were generally found for particles smaller than ~ 600
nm.
X
~
¦
~
X
~
X
¦
~
X
~
X
¦
A
X
~
X
¦
~
X
¦
~
i
I
<0.6 1.3-0.6 2.7-1.3 4.9-2.7
Size bin (nm)
~ HTC
¦ HTP
A LTC
X LTP
X LTIP
Source: Gutierrez-Daban et al. (2005).
Figure 3-40. Mass distributions for BaP at a high traffic urban center (HTC), high traffic urban
periphery (HTP), low traffic urban center (LTC), low traffic urban periphery (LTP), and
low traffic industrial urban periphery (LTIP) in Seville, Spain.
December 2008
3-84
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
+
~
o
X
S
O
+
+
+
S
~
+
+
A
~
1
X
i
A
~
$
*
X
o
A
~
f
X
X
A
2
~
B
*
i
0
¥
X
8
X
¦
X
¦
•
X
•
¦
¦
i
1.3-0.6 2.7-1.3 4.9-2.7
Size bin (^.m)
~ Naph
¦ Ace
A Acey
X Flu
X Phen
• Ant
+ Flua
-Pyr
-BaA
~ Chry
~ BbF
ABkF
x BaP
XlnP
ODbA
+ Bper
Source: Gutierrez-Da ban et al. (2005).
Figure 3-41. Mass distributions for sixteen PAHs at a high traffic city center in Seville, Spain.
Near roadway environments can exhibit particularly high concentration gradients. After initial
emission from a motor vehicle, the evolution of the PM distribution within the plume is a function of:
(1) the turbulence that dilutes the plume; and (2) evaporation or condensation of the volatile portion of the
aerosol. Figure 3-42 shows the size distribution measured by Zhu et al. (2002b) at distances of 17-300 m
away from the roadway (in this case, highway 710 in Los Angeles) and at an upwind site. It can be seen
that a mode originally measured around 9 nm increases in diameter and decreases in magnitude as
distance from the highway increases. Smaller secondary modes appear around 30 m from the roadway
with multiple modes at some particle sizes. By 150 m away from the highway, the size distribution
flattens with a small mode around 50 nm. It is clear from the figure that the fine particle sizes are better
represented by one sampler than ultrafine particles because the fine PMmass concentration is fairly
constant with distance up to at least 300 m. Zhou and Levy (2007) performed a meta-analysis of
traffic-related air pollution literature and found that levels of background pollution and meteorology can
have important impacts on the size of the elevated concentration region around the highway. Zhu et al.
(2002b) noted that small particles can be lost due to evaporation or to coagulation during Brownian
diffusion to form bigger particles. Zhang et al. (2005a) also saw this upward shift in mode diameter with
distance from the roadway in field measurements of ultrafine PM. The count distribution in the upper
ultrafine and fine PM ranges does not change drastically because the agglomeration process results in loss
December 2008
3-85
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
of particle number. Studies of particle sizes on roads (Kittelson et al., 2006b), in tunnels (Venkataraman et
al., 1994), and upwind and downwind of roads (Wilson and Suh, 1997; Zhu et al., 2002b) suggest that for
well-maintained spark-ignition vehicles most, and perhaps all, of the mass of particles emitted from the
vehicles are in the nuclei mode (i.e., smaller than accumulation mode). Traffic may also generate some
coarse mode particles (Wilson and Suh, 1997), presumably from material resuspended from the road. In
situations in which the dilution rates are lower than in a short tunnel or downwind of a road way,
condensation of condensable vapors can give rise to particles in the accumulation mode (Kittelson, 1998;
Wilson and Suh, 1997). Diesel engines, in particular, emit elemental (black) carbon in the lower end of
the accumulation mode, with number emissions dominated by semi-volatile material in the nuclei mode
(Kittelson, 1998; 2006a; 2006b). Sharp gradients in black carbon mass have been observed along
roadways with high diesel traffic (Zhu et al., 2002a). As the traffic pollution moves downwind, the
ultrafine particles may grow into the accumulation mode by coagulation or condensation. In addition to
Gaussian dispersion and wind eddies caused by the presence of natural and anthropogenic barriers,
Sahlodin et al. (2007) demonstrated that turbulence produced by vehicles can result in modification of the
plume emanating from the highway. Hence, on-road turbulence could potentially alter the aerosol size
distribution. This added turbulence could cause some evaporation of tiny nucleation particles that have
not absorbed or adsorbed onto soot nuclei.
Such knowledge of neighborhood-scale variability is important for interpreting data from PM10 and
PM2 5 community monitors. Figure 3-43 shows data derived from the fifteen CSAs/CBSAs for PM10 and
PM2 5 discussed in Section 3.5.1.2. This figure contains the inter-sampler correlations obtained for
sampler pairs located within a distance of 4 km (i.e. neighborhood scale). PM2 5 data appear to have a
flatter slope, with average correlation maintained at 93% within 4 km (R2 = 0.22). There is more scatter
and variability among the PMi0 data, with an average correlation of 70% within 4 km (R2 = 0.03). The
level of variability in PMi0 compared with PM2.5 relates to transport and dispersion of the PMi0.2 .5
component of PMi0 compared with PM2 5. However, differences in composition, source location,
topography, and monitor height—all of which could affect concentrations—could drive the relatively high
level of scatter for both size classes, considering the low computed R2 values for each of these curves.
December 2008
3-86
DRAFT-DO NOT QUOTE OR CITE
-------
17m.
J.Oe-5 •
C
JOe-5
ttt/1 !\
JO m ~ v M
I \
I Oe-J
I \
I ^ 150ra
Up Wind '--^Xy \ --100 m
0.8
0.6 -
10 |00
Particle Diameter, Dp (nm)
I
::
.6-25 i
04
25-50 nm
K 0J
00
50-100 nm
¦JJlT. — — — —J — — -J— <;t
100-220 urn
0 1WJ 2IXJ 300
Distance down wind from the 710 freeway (m)
Source: Zhu et al. (2002b).
Figure 3-42. Figure to be replaced. Particle size distributions measured at various distances from
the 710 freeway in Los Angeles (top), and particle number concentration as a function
of distance from the 710 freeway (bottom).
December 2008
3-87
DRAFT-DO NOT QUOTE OR CITE
-------
1.00
0.90
0.80
0.70
0.60
c
0
1 0.50
o
O
0.40
0.30
0.20
0.10
0.00
0 0.5 1 1.5 2 2.5 3 3.5 4
Distance Between Samplers (km)
Figure 3-43. Inter-sampler correlations as a function of distance between monitors for samplers
located within 4 km (neighborhood scale) for PM2.5 and PM10.
3.5.2. Temporal Variability
1 Temporal variability is another important factor in characterizing PM. This section addresses
2 trends, and seasonal and hourly variability. Trends in PMi0 and PM2.5 are addressed in Section 3.5.2.1
3 based on AQS data. Seasonality is coupled with spatial variability and has been discussed in the regional
4 context above. Section 3.5.2.2 below briefly investigates the seasonality on a finer time scale, thereby
5 addressing issues relating to the seasonal definitions used earlier. Section 3.5.2.3 addresses hourly
6 patterns, an issue particularly important to understanding the behavior of PM concentrations in reference
7 to sources, human activity patterns and exposure. Hourly patterns are investigated using AQS data on a
8 national basis for PM10 and PM2 5. Data for ultrafine particles are presented where available.
3.5.2.1. Trends
PM10
9 Figure 3-44 shows the 20-year trend in U.S. ambient 24-h PM]0 concentrations from 1988 to 2007
10 along with the number of monitoring sites above the NAAQS. In 2007, the U.S. national average second
"O» „ o
00 b" -.
o
* * - o
Rpm2.5 —0.0163D + 1 o
O R2 = 0.2178
- O. ° °
RpMio —0.0748D + 1
R2 = 0.0346
°
• PM2.5
o PM10
Linear (PM2.5)
- - Linear (PM 10)
December 2008
3-88
DRAFT-DO NOT QUOTE OR CITE
-------
1 highest PM10 concentration was 37 percent lower than in 1988 (see Figure 3-44a). Of 281 sites used in
2 this trend analysis, the number reporting concentrations above the level of the 24-h PM10 NAAQS
3 (150 |_ig/m ) fell from 23 in 1988 to 5 in 2007 with a max of 29 in 1989 (see Figure 3-44b). Figure 3-45
4 shows trends in the second highest 244i PM10 concentrations broken down by U.S. EPA region; all
5 regions saw an overall decrease from 1988 to 2007. Largest decreases occurred in EPA Region 10, which
6 incorporates Washington, Oregon, Idaho and Alaska. Most of the decrease occurred between 1988 and
7 1995.
"o>
0> CO
SI
<=> c
" o
-------
Q> o
Oi of!
Year
Coverage: 274 monitoring sites EPA Regions
in the EPA Regions (out of a total
of 879 sites measuring PM10 in ~ 0
2007) that have sufficient data to
assess PM10 trends since 1988.
Data source: U.S. EPA, 2008
Q
0
Source: U.S. EPA (2008a).
Figure 3-45. Ambient 24-h PM10 concentrations in the contiguous U.S. by EPA region, 1988-2007.
PM2.5
1 Figure 3-46 shows the trend in U.S. ambient 24-h PM2 5 concentrations from 1999 to 2007 along
2 with the number of monitoring sites above the 24-h NAAQS. In the period 2005-2007, the three-year
3 average of the 98th percentile of 24-h PM2 5 concentrations fell 10 percent from the 1999-2001 period
4 (see Figure 3-46a). The number of sites reporting values greater than the 24-h NAAQS declined 40
5 percent (see Figure 3-46b). Figure 3-47 illustrates the downward trend in the 98th percentile of 24-h PM2 ,
6 concentrations for three consecutive calendar years in all U.S. EPA regions. This trend is most
7 pronounced in Region 9 incorporating Arizona, California and Nevada where this value dropped 25%
8 from the 1999-2001 period to the 2005-2007 period.
—R1
— R2
— R3
— R4
—R5
R6
R7
R8
— R9
—R10
—Mat'l
NAAQS = 150 pg/m3
December 2008
3-90
DRAFT-DO NOT QUOTE OR CITE
-------
60
E 45
A. Ambient concentrations
¦S >>
30
15
0
90% of sites have
concentrations below this line
NAAQS = 35 MO/m3
i
' Median
A
Average
10% of sites have
concentrations below this iine
'99-01 '00-'02 '01-'03 02-04 '03-'05 '04-'06 '05-'07
Averaging period
400
B. Number of trend sites above NAAQS
ro CZ
"O fc
"8 300
Q3 ^ J.
>
GO
_ a
£ §
o OJ
200
o a> 100
C/5 OJ
E £ "99-01 '00-'02 '0t-'03 02-04 03-05 '04-'06 '05-'07
3 O
^ ° Averaging period
^Coverage: 718 monitoring sites in 489 counties nationwide (out of a
total of 831 sites measuring PM2.5 in 2007) that have sufficient data
to assess PM25 trends since 1999.
Source: U.S. EPA (2008a).
Figure 3-46. Ambient 24-h PM2.5 concentrations in the U.S., 1999-2007, showing (A) ambient
concentrations and (B) number of trends sites above the NAAQS.
December 2008
3-91
DRAFT-DO NOT QUOTE OR CITE
-------
05 '99-01 '00-'02 '01-'03 '02-04 '03-'05 '04-'06 '05-'07
Averaging period
NAAQS = 35 jjg/m
'Coverage: 697 monitoring sites
in the EPA Regions (out of a total
of 831 sites measuring PM2.5 in
2007) that have sufficient data to
assess PM2 5 trends since 1999.
Data source: U.S. EPA, 2008
EPA Regions
® © ©°
© ©0
\r
x-v
0
E
Source: U.S. EPA (2008a).
Figure 3-47. Ambient 24-h PM2.5 concentrations in the contiguous U.S. by EPA region, 1999-2007.
1 Figure 3-48 contains similar trend information for the annual PM2 5 NAAQS. The seasonally
2 weighted 3-y average PM2 5 concentrations for the years 2005 to 2007 were at the lowest since national
3 monitoring began in 1999 (see Figure 3-48a). The seasonally weighted 3-y average fell ten percent
4 between the 1999-2001 averaging period and the 2005-2007 averaging period. The number of sites
5 reporting concentrations above the annual average PM2.5 NAAQS fell 56 percent over these same periods
6 (see Figure 3-48b). Figure 3-49 illustrates the annual trends in PM25 by U.S. EPA region. Declines were
7 the greatest in Region 9 again where PM2 5 concentrations fell 20 percent from the 1999-2001 averaging
8 period to the 2005-2007 averaging period.
December 2008
3-92
DRAFT-DO NOT QUOTE OR CITE
-------
2 ~~
QJ •»— CT3
s =
so 22
c t= °
20
15
10
A. Ambient concentrations
90% of sites have concentrations below this line
NAAQS_=_
Median
*
Average
10% of sites have concentrations below this line
1 1 1 1 1
'99-'01 '00-'02 01-03 '02-04 '03-'05 '04-'06 '05-'07
Averaging period
B. Number of trend sites above NAAQS
OJ CO
*= 50
2 .ffi 100
99-01
0-03
02-04
•03-'05 '04-'06 '05-'07
Averaging period
Coverage: 718 monitoring sites in 489 counties nationwide (out of a
total of 802 sites measuring PM2 5 in 2007) that have sufficient data
to assess PM2.5 trends since 1999.
Source: U.S. EPA (2008a).
Figure 3-48. Ambient annual PM2.5 concentrations in the U.S., 1999-2007, showing (A) ambient
concentrations and (B) number of trends sites above the NAAQS.
December 2008
3-93
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
20
NAAQS = 15 pg/m:
— R1
—R2
— R3
— R4
—R5
— R6
R7
R8
—R9
—R10
—Nat'l
0| , , , , , L
99-01 '00-'02 '01-'03 02-04 03-05 04-'06 '05-07
Averaging period
'Coverage: 697 monitoring
EPA Regions
sites in the EPA Regions (out
of a total of 802 sites
P e.
o O
measuring PM2 5 in 2007) that
have sufficient data to assess
PM2.5 trends since 1999.
©
Data source: U.S. EPA, 2008
Source: U.S. EPA (2008a).
Figure 3-49. Ambient annual PM2.5 concentrations in the contiguous U.S. by EPA region, 1999-2007.
Several monitoring sites were excluded from the above trend analyses to provide a consistent basis
for comparison over the years of monitoring. This included exclusion of sites when there was no
corresponding site in later or earlier years.
Many of the figures and tables presented in the preceding sections have included a seasonal
break-down based on the following climatological seasons: winter (December-February), spring
(March-May), summer (June-August) and fall (September-November). Annex A contains plots of PM2.5
composition by individual month, illuminating intra-annual variability on a finer time scale. The same 15
CSAs/CBSAs are investigated and included in these plots; they are generated from the same data used in
the seasonal and annual pie charts based on the SANDWICH method shown earlier in Figures 3-14
through 3-18.
Monthly plots for most of the areas reveal heterogeneity in PM composition within the 3-month
long seasonal bins defined earlier. This is especially true in the spring and fall when concentrations are
changing rapidly with time. Many cities exhibit very rapid fluctuation in PM2 5 composition on short
timescales. For example, the NO;, mass in Los Angeles and Riverside can vary from a small fraction to
the most prevalent fraction of PM2 5 mass in a month's time based on the 3-y aggregate data in Annex A.
3.5.2.2. Seasonal Variations
December 2008
3-94
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
Therefore, selecting a different delineation point for the seasons can have an influence on the seasonal
composition analysis, specifically for constituents that fluctuate rapidly (e.g., NO, ).
Relatively little is known about the seasonal variability in ultrafine particles. Kuhn et al. (2005a)
and Zhu et al. (2004) found that the concentrations in the ultrafine mode can be much higher during
winter, particularly during evenings, because atmospheric dilution is reduced in response to lower mixing
heights. This can be seen in Figure 3-50. Mathis et al. (2005) found that emissions of particles in the
range of 45-900 nm are significantly higher with decreasing temperature and that cold-start conditions
produce roughly an order of magnitude greater PM number emissions in gasoline engines and more than
two orders of magnitude higher PM number emissions in diesel engines when compared with warm start
conditions.
200000
180000
„ 160000
g 140000
120000
100000
80000
60000
40000
20000
0
I
§
¦o
Site A Winter Even. -
Site A Winter Day -
Site A Summer -
20 30
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
3.5.2.3. Hourly Variability
Hourly PM10 and PM2 5 measurements are conducted at many sites using the beta gauge or TEOM.
Many of the hourly measurements for PM10 have FRM or FEM status. All available hourly data from
FRM, FEM and FRM-like monitors in the fifteen CSAs/CBSAs discussed earlier were used to investigate
diel variation in PM. Of the fifteen CSAs/CBSAs, Atlanta, Chicago, Pittsburgh, Seattle and St. Louis had
qualifying hourly PMi0 and PM2 5 data available. Denver, Detroit, Los Angeles, Philadelphia, Phoenix,
and Riverside had only qualifying hourly PMi0 data. Houston and New York had only qualifying PM2 5
data. Birmingham and Boston had no qualifying hourly PMi0 or PM2 5 data.
Annex A includes diel plots for PMi0 stratified by weekdays and weekends for eleven of the fifteen
CSAs/CBSAs with available data between 2005 and 2007. All cities show a gradual morning increase in
mean PMi0 starting at approximately 6:00 am on weekdays, corresponding with the start of the morning
rush hour before the break-up of overnight stagnation. The magnitude and duration of this peak, however,
varies considerably by area. Phoenix shows the most pronounced morning PMi0 peak concentration,
which drops off during the day and reappears in the evening. In contrast, Chicago shows a less
pronounced PMi0 peak concentration, which remains elevated throughout the day. Figure 3-51 shows the
diel plots of PM10 for Chicago and Phoenix. In both instances, the weekend diel pattern is similar in shape
to the weekday pattern with less exaggerated peaks.
December 2008
3-96 DRAFT-DO NOT QUOTE OR CITE
-------
Chicago
Weekday(N =1971)
291
218 -
145 -
73 -
0 -N*
Weekend (N = 793)
12 18 24
291 i
218 -
145
73
0 -F
¦ Median
Mean
90th & 10th
95th & 5th
12 18 24
Weekday (N = 1532)
Phoenix
291 ->
218 -
Weekend (N = 618)
Figure 3-51. Diel plot generated from hourly FEM PM10 data (ng/m3) stratified by weekday (left) and
weekend (right) for Chicago and Phoenix from 2005 to 2007. Included are the number
of monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for
each hour.
Annex A includes diel plots for PM2 5 stratified by weekdays and weekends for seven of the fifteen
CSAs/CBSAs with available data between 2005 and 2007. A similar morning PM2 5 peak starting at
approximately 6:00 am is present in all cities except Pittsburgh, where it appears that dispersion behavior
during the night results in elevated PM2 5 levels throughout the night that blend in with any morning peak.
With the exception of Pittsburgh, all seven metropolitan areas show two distinct daily peaks on both the
weekdays and weekends. The evening PM2 5 concentration peak is broader than the morning peak and
December 2008
3-97
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
extends to overnight hours, reflecting the concentration increase caused by a drop in boundary layer
height at night. Figure 3-52 compares the 2-peak shape of the PM2 5 diel distribution in Seattle with
Pittsburgh, where the overnight increase washes out the morning peak.
Pittsburgh
Weekday (N = 981) Weekend (N = 407)
77 v
77 t
csi
58 -
38
58 -
38 -
19
19 -
' I 1
6
¦ I -
12
¦ I "
18
"""I
24
6
12
¦ I "
18
24
Seattle
in
c\i
Weekday (N = 5775)
77 ~i
58 -
38 -
19
—r-
6
12 18
—i
24
Weekend (N = 2332)
77 ->
58 -
38 -
19 „
Median
Mean
90th & 10th
95th & 5th
I I
12 18
24
Figure 3-52. Diel plot generated from hourly FRM-like PM2.5 data (M.g/m3) stratified by weekday (left)
and weekend (right) for Pittsburgh and Seattle from 2005 to 2007. Included are the
number of monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles
for each hour.
December 2008
3-98
DRAFT-DO NOT QUOTE OR CITE
-------
Working days
Sundays
90
80
70
Q.
60
C.
50
X
/-s
40
z
30
20
10
0
- Jagtv.
-HC0
Diff Jgt.-HC0
rural
I
Im
Ii —«-
k Fl —*
y
• • •
» X I X 1-
4J
40000
35000
'•5
S3
c_
.u
30000
=tt
25000
_c
£
20000
E
15000
;z
3
©
e
10000
5000
Diff Jgt.-HC0
Jagtv
HC0
Diff Jgt-HC0
O
z
£
EL
O
u
it
40000
35000
c SS
30G0C
o. ^
— %
25000
° ji
20000
' E
15000
z 2
o u
10000
5000
0
Dff Jgt.-HC0
rura
DlfT.gt.-HC0
Diff Jgt.-HC0
-a
c3 . ,
Diff Jgt.-HC0
Diff Jgt.-HC0
Teom_Jgt
T eom_Jgt
Source: Ketzel et al. (2003).
Figure 3-53. Average diurnal variation of NOx, CO, particle number and particle volume on
weekdays (left) and Sundays (right).
1 Ultrafine particles in urban environments have been shown to exhibit a similar two-peaked diel
2 pattern in Los Angeles (Moore et al., 2007; Sardar et al., 2005) and the San Joaquin Valley (Herner et al.,
3 2005) in California as well as in Kawasaki City, Japan (Hasegawa et al., 2005) and Copenhagen,
4 Denmark (Ketzel et al., 2003). Figure 3-53 from the Denmark study shows a large peak in total particle
5 number (dominated by ultrafine particles) corresponding with the morning rush hour. The morning peak
December 2008
3-99
DRAFT-DO NOT QUOTE OR CITE
-------
is absent on Sundays, however. Many studies also show a broad afternoon ultrafine concentration peak,
which likely originates from a combination of primary sources such as evening rush hour traffic and
secondary formation of ultrafine particles in the atmosphere. Secondary formation likely contributes a
substantial amount of ultrafine particles since the evening peak is present on weekends whereas the
morning traffic related peak essentially vanishes.
Winter
Spring
PM2.5 (daily avg)-
PM10-2.5 (daily avg)«
SO 2 (daily avg)-
N02 (daily avg)"
CO (daily avg)"
03 (daily max 8-hr)-
~!—I—I—I—I—I—I—i—T
mm m ii hi ****
i—i—i—i—i—i—i—i—r
•
* ***#M
1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1
Summer
Fall
PM2.5 (daily avg)"
PM10-2.5 (daily avg)«
SO 2 (daily avg)-
N02 (daily avg)-
CO (daily avg)-
03 (daily max 8-hr)-
*
i
*
*
—i 1—i—i 1—i—i—i—r
i—i—i—i—i—i—i—i—i—
** * **#* ** i mm
iMWMHMMHW#*#
—i 1—i—i—i 1—i—i—r
i—i—i—i—i—i—i—i—i—
r (correlation coefficient)
r (correlation coefficient)
Figure 3-54. Correlations between 24-h PM10 and co-located 24-h average PM2.5, PM10-2.5, SO2, NO2
and CO and daily max 8-h avg O3 for the U.S. stratified by season (2005-2007). One
point is included for each monitor pair with the mean (green stars) and median (red
squares) of all correlations superimposed.
3.5.3. Statistical Associations with Copollutants
Associations between PM and other copollutants including S02, N02, CO and O , are investigated
in this section. AQS data were obtained from all available co-located monitors across the U.S. after
application of an 11 or more observations per quarter completeness criteria. Pearson correlation
December 2008
3-100
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
coefficients (R) were calculated using 2005-2007 data. The results are displayed graphically in Figure 3-
54 for correlations with PM10 and Figure 3-55 for correlations with PM2 5.
PM10 (daily avg)
PM10-2.5 (daily avg)
SO 2 (daily avg)
N02 (daily avg)
CO (daily avg)
03 (daily max 8-hr)
Winter
Spring
* * * » *#•****# m
~i—i—i—i—i—i—r
n—i—i—i—r
n—i—i—i—i—i—i—i—r
n—i—i—i—i—i—r
Summer
Fall
PM10 (daily avg)*
PM 10-2.5 (daily avg)-
SO 2 (daily avg)«
N02 (daily avg)«
CO (daily avg)-
03 (daily max 8-hr)-
—i 1—i—i 1—i—r~
~i—i—i—i—i—
—i 1—i—i—i 1—i—
~i—i—i—i—i—i—i—
r (correlation coefficient)
r (correlation coefficient)
Figure 3-55. Correlations between 24-h PM2.5 and co-located 24-h average PM10, PM10 2.5, SO2, NO2
and CO and daily max 8-h avg O3 for the U.S. stratified by season (2005-2007). One
point is included for each monitor pair with the mean (green stars) and median (red
squares) of all correlations superimposed.
For both PM10 and PM2 5 national composite copollutant correlations, there is considerable spread
in the observed correlations in all four seasons. On average, PM10 and PM2 5 correlate with each other
better than with the gaseous copollutants. The correlations between PM10 and PM2 5 are all positive but
span the range from just above zero to near one. This illustrates the wide variability in correlation
between these two PM metrics. Fewer points are available for correlation with PM10_2 5 because only data
from low-volume FRM/FRM-like samplers were used to calculate PMi0.2.5. The available data suggest a
stronger correlation between PMi0 and PMi0.2.5 than between PM2 5 and PMi0.2.5 on a national basis.
The correlation between PM and the gaseous pollutants included in Figures 3-54 and 3-55 also
have a large range in values using the national composite data. There is little seasonal variability in the
December 2008
3-101
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
mean correlation between PM and S02. N02 and CO, however, show higher correlations with PM on
average in the wintertime than in the other seasons. This is possibly driven by meteorology with increased
frequency of stagnation events in colder months. The correlation between daily max 8-h average 03 and
PM shows high seasonal variability with positive correlations on average in the spring, summer and fall
and negative correlations on average in the winter. The highest positive correlations are in the summer,
likely driven by favorable photochemical formation conditions for both 03 and secondary aerosols
(Joseph, 2008; Meng et al., 1997). The mean correlation drops below zero (-0.2 for PMi0 and -0.3 for
PM25) in the wintertime. As discussed in chapter 3 of the last AQCD for Ozone and other Photochemical
Oxidants (U.S. EPA, 2006c), this situation arises because photochemical production of ozone in the
planetary boundary layer is much smaller during the winter than summer, while primary PM levels are
elevated in many areas as a product of heating emissions and lower mixing heights. Ozone in the
boundary layer is mainly associated with the subsidence from above the boundary layer following the
passage of cold fronts. This subsiding air has much lower PM concentrations than were present in the
boundary layer. Therefore, a negative association between 03 and PM2 5 is frequently observed in the
wintertime. Bell et al. (2007a) also observed a wintertime minima in same-day correlations between 24-h
average PM and 03 using data from 98 U.S. urban communities over a 14-year period (1987-2000). The
average correlations were not negative in wintertime, however, as seen here. Furthermore, the highest
national average correlations were in spring and fall in the Bell et al. (2007a) analysis rather than summer
as observed here. This discrepancy could be a result of the different averaging times used for 03 or the
selection of different monitoring networks and/or time periods.
Correlations among copollutants for individual CSAs/CBSAs are included in Annex A for PM10
and Annex A for PM2 5. The same fifteen CSAs/CBSAs were chosen for further investigation, but several
had an insufficient amount of co-located data to be included. As can be seen from the individual
CSAs/CBSAs in these figures with multiple pairs of co-located monitors per pollutant, there can be
considerable variation in the correlations even on an urban scale. Birmingham, Boston, and St. Louis all
show positive wintertime correlations between PMi0 and daily maximum 8-h average 03; Denver, Detroit,
Houston, Los Angeles and Phoenix show negative wintertime correlations. The remaining seven
CSAs/CBSAs have insufficient data. For PM2 5, all selected cities with sufficient data show negative
correlations in the wintertime with daily max 8-h average 03 (including Birmingham, Boston, Chicago,
Denver, Houston, Los Angeles, Philadelphia, Phoenix, Pittsburgh, Riverside and St. Louis). The
remaining four CSAs/CBSAs have insufficient data. In Baltimore (not one of the fifteen CSAs/CBSAs
included in this investigation), Sarnat et al. (2001) found a significant (at the p <0.05 level) positive (0.67)
and negative (-0.72) correlation between daily PM2 5 and 03 in the summer (June 19-August 23, 1998)
and winter (February 2-March 13, 1999), respectively. The negative correlation between PM2 5 and 03
December 2008
3-102
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
observed in many cities across the U.S. in the wintertime illustrates the importance of considering
seasonality when assessing correlations between these air pollutants. While not as extreme, the other
pollutants can exhibit seasonal variation in correlations with PM as well.
3.5.4. Estimating Source Contributions to PM
Methods for analyzing the composition of ambient PM samples in terms of contributions from
different sources are reviewed in this section. Associations between exposures to ambient PM matter, as
represented by ambient monitors, and health outcomes have been extensively studied. Some health
studies, described in Section 6.6, have used source apportionment methods to evaluate relationships
between health outcomes and PM (mainly PM2 5) from different sources. This section is intended to aid
interpretation of those study results. Understanding the contribution of different emissions sources to
ambient PM is also important in evaluating air quality data.
3.5.4.1. Receptor Models
Receptor models are diagnostic in their approach (i.e., they attempt to derive categories of source
contributions based either on ambient data alone or in combination with data on the chemical composition
of sources). Their formulation contrasts with that of other deterministic models (i.e., three-dimensional
chemistry and transport models) that are formulated in a prognostic, or predictive manner (i.e., they
attempt to predict species concentrations using a mass or species conservation equation that includes
terms based on emissions inventories, meteorology, atmospheric transport, chemical transformations, and
deposition). Receptor models have been primarily employed as part of the development of air quality
management plans. However, there have been several publications relating apportioned source types
based on receptor models to human health effects. Discussions in this section will focus mainly on those
methods that have been used to relate health outcomes to sources. More complete descriptions of a large
number of types of receptor models currently in use are given in Watson et al. (2008) who summarize the
properties of these methods, including the strengths and weaknesses. This compilation of receptor models,
broken down into different approaches (i.e., chemical mass balance, factor analysis, tracer-based,
meteorology based) is in Tables A-X through A-X.
Receptor models such as the chemical mass balance (CMB) model (Watson et al., 1990) relate
source category contributions to ambient PM concentrations based on analyses of the compositional
profiles of ambient and source emissions samples. It uses as its basis a mass balance equation that
represents all chemical species in an aerosol sample as linear combinations of contributions from a fixed
December 2008
3-103
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
number of independent sources plus an error term representing the portion of the measurement that cannot
be fit by the model.
The compositional profiles used in receptor models can be extensive. As an example, several
studies have identified EC and over 100 organic carbon compounds in gasoline PM emissions, including
alkanes, PAHs, oxy-PAHs, steranes, hopanes, and organic acids (Schauer et al., 1999, 2002). This
breakdown in identifiable groups of organic compounds is illustrated in Figure 3-56 and Table 3-20.
Maricq (2007) noted that these 100+ compounds constituted less than 5% of total organic compounds in
the PM samples analyzed. Geller et al. (2006) noted that emissions factors for PAHs during steady-speed
operation were 40-60% lower in diesel vehicles and more than 90% lower in gasoline vehicles during
fluctuating speed operation. Riddle et al. (2007) found that low emissions gasoline vehicles emitted more
low molecular weight PAHs but fewer high molecular weight and large PAHs when compared with
gasoline vehicles employing three-way catalysts for hydrocarbon control. This was also observed by
Fraser et al. (1999) and Schauer et al. (2002). Additionally, several trace metals, including Al, Ca, Fe, K,
Mg, Na, Ba, Cr, Cu, Mn, Ni, Pb, S, Ti, V, and Zn have been identified in gasoline and diesel PM
emissions in significantly higher amounts for variable speed operation than under steady operation. Geller
et al. (2006) also tested the reduction-oxidation potential of ultrafine and accumulation mode PM
generated from diesel and gasoline under variable speed and steady speed conditions and found that diesel
PM under transient conditions was approximately 14% greater than gasoline PM under transient
conditions and 56% greater than diesel PM under steady driving conditions. Escribano et al. (2001) and
Sadezky et al. (2005) compared the diesel PM Raman spectroscopy curve with that of graphite and found
peaks at similar wave numbers. Data for the compositional profiles for several other important sources of
PM that could be used for CMB modeling are shown in Annex A.
December 2008
3-104
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
-Oxy-PAH
Steranes
^Hoparies
Source: Fraser et al. (1999).
Figure 3-56. Schematic of organic composition of particulate emissions from gasoline-fueled
vehicles.
In other methods, various forms of factor analysis are used that rely on the varying mix of species
present in ambient observations of compositional data to derive the source contributions. Standard factor
analytic approaches such as Principal Component Analysis (PCA) have been used but PCA alone can
apportion only the variance, not the mass, in an aerosol composition data set. Additional steps such as
those applied in Absolute Principal Components Scores (APCS) are required to apportion mass from PCA
(Miller et al., 2002; Thurston and Spengler, 1985). In Positive Matrix Factorization (PMF) (Paatero and
Tapper, 1994), which is becoming more widespread in its use, the ambient compositional data matrix is
decomposed into the product of a matrix representing the source contributions and one representing the
source profiles. Solutions are obtained by minimizing an object function with respect to these two
matrices, and solutions are subject to non-negativity constraints. PMF also allows for the treatment of
missing data and data near or below detection limits by weighting elements inversely according to their
uncertainties. The PMF approach can only be applied to time series data, whereas CMB can be applied to
one sample. Both the CMB and the PMF approaches find solutions based on least squares fitting and
minimization of an object function. Both methods provide error estimates for the solutions based on
estimates of the errors in the input parameters. It should be remembered, though, that the error estimates
for both methods often contain subjective judgments about the magnitude of the analytical and monitoring
errors.
Fine Organic Extractable and Resolved Identified
Compounds Elutable Organics Organics
142 mg/l 118mg/l 18 mg/l 5 mgyi
Organic Acids
Alkanes
PAH
Not Extractable
or Elutable
Extractable
and Elutable
Unidentified
Identified
Unresolved
Resolved
December 2008
3-105
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Table 3-20.
Emissions factors (ng/kg) for trace elements under variable speed and steady speed
driving conditions for PM emitted by diesel and gasoline engines.
Diesel
Gasoline
Transient
Steady State
Transient
Steady State
Al
9108(5224)
2706
2273(545)
252
Ca
69,443(23,640)
16,128
18,247(3044)
2324
Fe
22,910(21,448)
2036
10,266(9928)
138
K
4672(752)
1191
1935(558)
117
Mg
3087(461
997
5183(1706)
183
Na
7736(1751)
1945
2237(1125)
321
Ba
583(349)
73
331 (55)
4.8
Be
26(12)
23
6.7(1.1)
1.5
Cr
634(354)
93
138(6.7)
8.6
Cu
1944(679)
627
1745(1803)
16
Li
13(02)
7.9
3.0(1.4)
0.9
Mn
368(183)
76
152(85)
3.4
Ni
2310(656)
644
107(0.7)
21
Pb
793(593)
79
237(2.3)
11
S
23,750(5295)
6713
8705(3375)
349
Ti
2036(320)
345
118(9.3)
24
V
28(94)
11
15(11)
1.8
Zn
21,118(4422)
5620
4650(1225)
198
Source: Gelleretal. (2006).
The nature of the solutions in terms of source categories is different in the CMB and PMF
approaches. In the CMB approach, the composition of the source emissions is assumed to be known
based on measurements. These assumptions may or may not reflect the composition of emissions
affecting a particular site at any given time or place. However, there may be variations in the composition
of individual source categories (e.g., soils, motor vehicle emissions) across a given airshed and even in
the composition of the same source with time. Source profiles can also be altered between emission and
receptor locations resulting from atmospheric reactions, depending on the source type and species under
analysis. The CMB technique was developed for apportioning source categories of primary PM and was
not formulated to include sources of secondary PM. All of the mass cannot be apportioned unless there is
information for the composition of all sources affecting a given site.
In PMF, the source solutions are more general in that they contain information about the
entrainment of emissions from additional sources during transport, the time dependence of the
composition of emissions from particular sources, the formation of secondary species and local
December 2008
3-106
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
differences in source compositions. PMF differs from CMB because it derives the source mix from
measured data. However, the procedure used to find a solution results in some rotational ambiguity
(Paatero and Tapper, 1994). The assignment of sources and the factors modifying them depend largely on
past experience and judgments based on data for source profiles to identify the sources and their
modifying factors contained in the solution. These issues are alleviated to some degree by folding in
information about local wind fields and other physical parameters.
The UNMIX model takes a geometric approach that exploits the covariance of the ambient data to
determine the number of sources, the composition and contributions of the sources, and the uncertainties
(Henry, 1997). UNMIX uses PCAto find edges in m-dimensional space, where m is the number of
ambient species. Success of the UNMIX model hinges on the ability to find these "edges" in the ambient
data from which the number of source types and the source compositions are extracted. In simplest terms,
the approach can be seen to be similar to that for deriving ternary mixing diagrams, except there is
extension to higher dimensionality. Measurement errors in the ambient data "fuzz" the edges, making
them difficult to find. UNMIX employs an "edge-finding" algorithm to find the best edges in the presence
of error. UNMIX does not make explicit use of errors or uncertainties in the ambient concentrations,
unlike the methods outlined above. Rather they are implicitly incorporated into the analyses.
Partial least squares (PLS) is another mathematical model related to PCA which has been used in a
limited number of PM toxicology studies to establish a relationship between PM constituents and health
outcomes (McDonald et al., 2004; Seagrave et al., 2006; Veranth et al., 2006). Unlike PCA and other
receptor models discussed in this section, PLS incorporates both predictor variables (e.g., PM component
concentrations) and outcome variables (e.g., toxicological responses) into one coupled regression model
rather than relying on a stand-alone regression analysis after the fact. Like PCA, PLS groups the
observable variables into a reduced number of latent variables, thereby reducing the dimensionality of the
model (and, hence, PLS also stands for "projection to latent structures"). Typically, PM toxicology studies
have been limited to two-component models (two latent variables on the predictor side compared with
two on the outcome side), thereby producing a 2x2 loading plot revealing relationships between
predictors and outcomes. PLS is particularly useful when there are more predictor variables than
observations, which is a situation that other multivariate factor analysis approaches do not handle well.
However, since PLS is a variance based approach, it shares the same shortcomings discussed earlier for
PCA. PLS has also traditionally been limited to two-component applications even though this is not a
strict mathematical limitation.
December 2008
3-107
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Results from Receptor Models
Results of receptor modeling calculations indicate that PM2 5 is most often produced mainly by
fossil fuel combustion. Fugitive dust, found mainly in the PM10_2 5 size range, represents the largest source
of measured ambient PM10 in many locations in the western U.S. Quoted uncertainties in the source
apportionment of constituents in ambient aerosol samples typically range from 10 to 50%. It is apparent
that a relatively small number of broadly defined source categories, compared to the total number of
chemical species that typically are measured in ambient monitoring-source receptor model studies, are
needed to account for the majority of the observed mass of PM in these studies. Compilations of source
attribution studies using CMB have appeared in the PM AQCD (U.S. EPA, 2004) for PMi0 and using
PMF in Engel-Cox and Weber (2007). Results of the compilation by Engel-Cox and Weber (2007) for the
eastern U.S. are shown in Figure 3-57. There are only three main source categories in the figure
constituting most of the PM2 5 mass. Two of these are predominantly secondary and not identified by
sources of precursors. Tables listing results of other receptor modeling studies for PM2 5 and PMi0, many
of which are in the western U.S., are given in Annex A.
~3
TOTAL Sulfate Nitrate Mobile Luriiius; Industrial Crustal&Salt Other
Source
Source: Engel-Cox and Weber (2007)
Figure 3-57 Source category contributions to PM2.5 at a number of sites in the East derived using
PMF.
December 2008
3-108
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
Spaf/a/ Variability in Source Contributions to PM Based on Receptor Models
Spatial variability in source contributions across urban areas is an important consideration in
assessing the likelihood of exposure measurement error in epidemiologic studies relating health endpoints
to sources. Arguments similar to those for using ambient concentrations as surrogates for personal
exposures apply here. Studies for PM2 5 (Kim et al., 2005b; Wongphatarakul et al., 1998) indicate that
intra-urban variability increases in the following order: regional (e.g., secondary S042 from EGUs) < area
(on-road mobile sources) < point (stacks) sources. This point is illustrated in Figure 3-58. The only study
available for PMi0.2.5 (Hwang et al., 2008) indicates a similar ordering, but without a regional component
(resulting from the short lifetime of coarse PM compared to transport times on the regional scale) as
shown in Figure 3-59.
O 0.6
© 0.4
Source: Kim et al. (2005b)
Figure 3-58. Pearson correlation coefficients for source category contributions to PM2.5 at the ten
Regional Air Pollution Study/Regional Air Monitoring System (RAPS/RAMS)
monitoring sites in St. Louis.
December 2008
3-109
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
0.8
0.6 -
Source: Hwang et al. (2008)
Figure 3-59. Pearson correlation coefficients for source contributions to PM10-2.5 at the ten
Regional Air Pollution Study/Regional Air Monitoring System (RAPS/RAMS)
monitoring sites in St. Louis.
Chemical Transport Models
CTMs are the prime tools used to compute the interactions among atmospheric pollutants and their
transformation products, the production of secondary aerosols, the evolution of particle size distribution,
and transport and deposition of pollutants. CTMs are driven by emissions inventories for primary species
such as NOx, SOx, NH3, and primary PM, and by meteorological fields produced by other numerical
prediction models. Meteorological quantities such as winds and temperatures are taken from operational
analyses, reanalyses, or weather circulation models. In most cases, these are off-line meteorological
analyses, meaning that they are not modified by radiatively active species generated by the air quality
model (AQM), although new, integrated models of meteorology and chemistry are now available as well
(e.g., Binkowski et al., 2007).
Emissions of precursor compounds can be divided into anthropogenic and biogenic source
categories, and biogenic sources can be further divided into biotic (vegetation, microbes, animals) and
December 2008
3-110
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
abiotic (biomass burning, lightning, geogenic) categories as presented above. However, the distinction
between biogenic sources and anthropogenic sources is often difficult to make, as human activities affect
directly or indirectly emissions from what would have been considered biogenic sources during the
preindustrial era. Thus, emissions from plants and animals used in agriculture have been referred to as
anthropogenic or biogenic in different applications. Wildfire emissions may be considered to be biogenic,
except that forest management practices may have led to the buildup of fuels on the forest floor, thereby
altering the frequency and severity of forest fires.
The initial conditions, or starting concentration fields of all species computed by a model, and the
boundary conditions, or concentrations of species along the horizontal and upper boundaries of the model
domain throughout the simulation, must be specified at the beginning of the simulation. Both initial and
boundary conditions can be estimated from models or data or, more generally, model + data hybrids.
Because data for vertical profiles of most species of interest are sparse, results of model simulations over
larger, usually global, domains are often used. As might be expected, the influence of boundary conditions
depends on the lifetime of the species under consideration and the time scales for transport from the
boundaries to the interior of the model.
Each of the model components described above has associated uncertainties and the relative
importance of these uncertainties varies with the modeling application. The largest errors in
photochemical modeling are still thought to arise from the meteorological and emissions inputs to the
model (Russell and Dennis, 2000). While the effects of poorly specified boundary conditions propagate
through the model's domain, the effects of these errors remain undetermined. Because many
meteorological processes occur on spatial scales smaller than the model grid spacing (either horizontally
or vertically) and thus are not calculated explicitly, parameterizations of these processes must be used.
These introduce additional uncertainty. Because the chemical production and loss terms in the continuity
equations for individual species are coupled, the chemical calculations must be performed iteratively until
calculated concentrations converge to within some preset criterion. The number of iterations and the
convergence criteria chosen also can introduce error.
CTMs have been developed for application over a wide range of spatial scales ranging up from
neighborhood to global. CTMs are used to: (1) obtain better understanding of the processes controlling
the formation, transport, and destruction of gas- and particle-phase criteria and hazardous air pollutants;
(2) understand the relations between concentrations of secondary pollutant products and concentrations of
their precursors; (3) understand relations among the concentration patterns of various pollutants that may
exert adverse effects; and (4) evaluate how changes in emissions propagate through the atmospheric
system to secondary products and deposition. More detailed discussion of CTM applications appears in
the 2008 ISA for NOx and SOx - Ecological Criteria (U.S. EPA, 2008e).
December 2008
3-111
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Global Scale
Global-scale CTMs are used to address issues associated with climate change and stratospheric 03
depletion, and to provide boundary conditions for the regional-scale models. The CTMs include
simplified mathematical descriptions of atmospheric transport, the transfer of solar radiation through the
atmosphere, chemical reactions, and removal to the surface by turbulent motions and precipitation for
pollutants emitted into the model domain. The upper boundaries of the CTMs extend anywhere from the
top of the mixed layer to the mesopause at -80 km in order to obtain more realistic boundary conditions
for problems involving stratospheric dynamics.
Global simulations are typically conducted at a horizontal resolution of 200 km2 or more.
Simulations of the effects of transport from long-range transport link multiple horizontal resolutions from
the global to the local scale. Finer resolution will only improve scientific understanding to the extent that
the governing processes are more accurately described at that scale. Consequently, there is a critical need
for observations at the appropriate scales to evaluate the scientific understanding represented by the
models.
Regional Scale
Most major modeling efforts in the EPA center on the Community Multiscale Air Quality modeling
system (CMAQ) (2006; Byun and Ching, 1999). [A number of other modeling platforms using
Lagrangian and Eulerian frameworks were reviewed in the 2006 AQCD for 03 (U.S. EPA, 2006c) and in
Russell and Dennis (2000)]. The capabilities of a number of CTMs designed to study local- and regional-
scale air pollution problems were summarized by Russell and Dennis (2000). Evaluations of the
performance of CMAQ are given in Arnold et al. (2003), Eder and Yu (2006), Appel et al. (2005), and
Fuentes and Raftery (2005). CMAQ's horizontal domain can extend from a few square kilometers to the
entire hemisphere. In addition, both of these classes of models allow resolution of the calculations over
specified areas to vary. CMAQ is most often driven by the MM5 mesoscale meteorological model
(Seaman, 2000), though it may be driven by other meteorological models including WRF and RAMS.
Simulations of pollution episodes over regional domains have been performed with a horizontal
resolution as low as 1 km, and smaller applications over limited domains have been performed at even
finer scales. However, simulations at such high resolutions require better parameterizations of
meteorological processes such as boundary layer fluxes, deep convection and clouds (Seaman, 2000), as
well as finer-scale emissions than are generally available at present. Finer spatial resolution is necessary
to resolve features such as urban heat island circulation; sea, bay, and land breezes; mountain and valley
breezes; and the nocturnal low-level jet, all of which can affect pollutant concentrations.
December 2008
3-112
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
The vertical resolution of these CTMs is variable, and usually configured to have more layers in the
PBL and fewer higher up. Because the height of the boundary layer is of critical importance in
simulations of air quality, improved resolution of the boundary layer height would likely improve air
quality simulations. Additionally, current CTMs do not adequately resolve fine scale features such as the
nocturnal low-level jet in part because little is known about the nighttime boundary layer.
CTMs require time-dependent, three-dimensional wind fields for the period of simulation. The
winds may be generated either by a model using initial fields alone or with four-dimensional data
assimilation to improve the model's performance; i.e., model equations can be updated periodically to
bring results into agreement with observations. Modeling efforts typically focus on simulations of several
days' duration, the typical time scale for individual 03 episodes. Longer term modeling series of several
months or multiple seasons of the year are now common. The current trend in modeling applications is
towards annual simulations. This trend is driven in part by the need to improve understanding of
observations of periods of high wintertime PM (e.g., Blanchard et al., 2002) and the need to simulate 03
episodes occurring in spring, fall, and winter.
Chemical kinetics mechanisms (sets of chemical reactions) representing the important reactions
occurring in the atmosphere are used in CTMs to estimate the rates of chemical formation and destruction
of each pollutant simulated as a function of time. Unfortunately, chemical mechanisms that explicitly treat
the reactions of individual reactive species are too computationally demanding to be incorporated into
CTMs for regulatory use. So, for example, are very extensive "master mechanisms" (Derwent et al.,
2001) that include approximately 10,500 reactions involving 3603 chemical species (Derwent et al., 2001)
combined into mechanisms that group compounds of similar chemistry together.
CMAQ and other state-of-the-science CTMs incorporate processes and interactions of aerosol-
phase chemistry (Mebust et al., 2003). There have also been several attempts to study the feedbacks of
chemistry on atmospheric dynamics using meteorological models, like MM5 for example (Grell et al.,
2000; Liu et al., 2001; Lu et al., 1997; Park et al., 2001). This coupling is necessary to accurately simulate
feedbacks which may be caused by the heavy aerosol loading found in forest fire plumes (Lu et al., 1997;
Park et al., 2001) or in heavily polluted areas. Photolysis rates in CMAQ can now be calculated
interactively with model produced 03, N02, and aerosol fields (Binkowski et al., 2007).
Spatial and temporal characterizations of anthropogenic and biogenic precursor emissions must be
specified as inputs to a CTM. Emissions inventories have been compiled on grids of varying resolution
for many HCs, aldehydes, ketones, CO, NH3, and NOx. Emissions inventories for many species require
the application of algorithms for calculating the dependence of emissions on physical variables such as
temperature and to convert the inventories into formatted emission files which can be used by a CTM. For
example, preprocessing of emissions data for CMAQ is done by the Spare-Matrix Operator Kernel
December 2008
3-113
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Emissions (SMOKE) system (http://smoke-model.oni). For many species, information concerning the
temporal variability of emissions is lacking, so long-term annual averages are used in short-term, episodic
simulations. Annual emissions estimates are often modified by the emissions model to produce emissions
more characteristic of the time of day and season. Significant errors in emissions can occur if
inappropriate time dependence is used. Additional complexity arises in model calculations because
different chemical mechanisms can include different species, and inventories constructed for use with
another mechanism must be adjusted to reflect these differences. This problem also complicates
comparisons of the outputs of these models because one chemical mechanism may produce some species
not present in another mechanism yet neither prediction may agree with the measurements.
Sub-Regional Scale
The grid spacing in regional CTMs, usually between 1 and 12 km2, is usually too coarse to resolve
spatial variations on the neighborhood scale. The interface between regional scale models and models of
smaller exposure scales described is provided by smaller scale dispersion models. Several models could
be used to simulate concentration fields near roads, each with its own set of strengths and weaknesses.
For example, AERMOD (http://www.epa.gov/scramOO 1/dispersion prcfrcc.htm) is a steady-state plume
model formulated as a replacement to the ISC3 dispersion model. In the stable boundary layer (SBL), it
assumes the concentration distribution to be Gaussian in both the vertical and horizontal dimensions. In
the convective boundary layer, the horizontal distribution is also assumed to be Gaussian, but the vertical
distribution is described with a bi-Gaussian probability density function (pdf). AERMOD has provisions
to be applied to flat and complex terrain and multiple source types (including, point, area and volume
sources) in both urban and rural areas. It incorporates air dispersion based on PBL turbulence structure
and scaling concepts and is meant to treat both surface and elevated sources and simple and complex
terrain in rural and urban areas. The dispersion of emissions from line sources like highways is treated as
the sum of emissions from a number of point sources placed side by side. However, emissions are usually
not in steady state and there are different functional relationships between buoyant plume rise in point and
line sources. However, AERMOD does not have provision for including secondary sources.
There are non-steady state models that incorporate plume rise explicitly from different types of
sources. For example, CALPUFF (http://www.src.com/calpuff/calpuff 1 .htm) is a non-steady-state puff
dispersion model that simulates the effects of time- and space-varying meteorological conditions on
pollution transport, transformation, and removal and has provisions for calculating dispersion from
surface sources. However, it should be noted that neither CALPUFF nor AERMOD was designed to treat
the dispersion of emissions from roads or to include secondary sources. In using either model, the user
would have to specify dispersion parameters that are specific to traffic. The distinction between a steady-
December 2008
3-114
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
state and time varying model might not be important for long time scales; however for short time scales,
the temporal variability in traffic emissions could result in underestimation of peak concentration and
exposures.
3.6. Background PM
The background concentrations of PM useful for risk and policy assessments informing decisions
about NAAQS are referred to as Policy Relevant Background (PRB) concentrations and are those
concentrations that would occur in the U.S. in the absence of anthropogenic emissions in continental
North America (defined here as the U.S., Canada, and Mexico). PRB concentrations include contributions
from natural sources everywhere in the world and from anthropogenic sources outside these three
countries. Background levels so defined facilitate separation of pollution levels that can be controlled by
U.S. regulations (or through international agreements with neighboring countries) from levels that are
generally uncontrollable by the U.S. These levels can also be used in quantitative risk assessments of
human health and environmental effects. In this section, estimates are provided for daily average and
annual average PRB concentrations are needed, corresponding to averaging times for PM NAAQS.
3.6.1. Contributors to PRB levels of PM
Contributions to PRB levels of PM include both primary and secondary natural and anthropogenic
components. Natural sources include wind erosion of natural surfaces (Gillette and Hanson, 1989);
volcanic production of S042 , primary biological aerosol particles; wild fires producing EC, OC, and
inorganic and organic PM precursors; and SOA produced by oxidation of biogenic hydrocarbons such as
isoprene and terpenes. However, human intervention can be involved in the formation of SOA, as
production of natural SOA depends to a large extent on the presence of anthropogenic NOx. As described
earlier in Section 3.3, prescribed fires are considered as part of PRB. In addition to emissions from forest
fires in the U.S., emissions from forest fires in other countries can be transported to the U.S. Boreal forest
fires in Canada (e.g., Mathur, 2008) and Siberia (Generoso et al., 2007) and tropical forest fires in the
Yucatan Peninsula and Central America (e.g., Wang et al., 2006a) have affected PM levels in the U.S.
PRB PM varies across the CONUS by region and season as a function of complex mechanism of
transport, dispersion, deposition, and reentrainment.
Dust from the Sahara desert and the Sahel in North Africa (e.g., Chiapello et al., 2005) affects
mainly the eastern U.S.; dust from the Gobi and Taklimikan deserts in Asia (e.g., VanCuren and Cahill,
2002; Yu et al., 2008) have the largest effects in the western U.S. but also affect air quality in the eastern
December 2008
3-115
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
U.S. Husar et al. (2001) report that the average PM10 level at over 150 reporting stations throughout the
northwestern U.S. was 65 |ig/m3 during an episode in the last week in April 1998, compared to an average
of about 20 |ig/m3 during the rest of April and May.
PRB contributions to PM2 5, PM10-2.5, and PMi0 can also be viewed as coming from two
conceptually separate components a reasonably consistent "baseline" component and an episodic
component. The baseline component consists of contributions that are generally well characterized by a
reasonably consistent distribution of daily values each year, although there is variability by region and
season. The episodic component consists of infrequent, sporadic contributions from natural
high-concentration events occurring over shorter periods of time (e.g. hours to several days) both within
North America (e.g. volcanic eruptions, large forest fires, dust storms) and outside North America (e.g.,
transport related to dust storms from deserts in North Africa and China and storms at sea). These episodic
natural events, as well as events like the uncontrolled biomass burning in Central America, are essentially
uncontrollable and do not necessarily occur in all years.
In-situ measurements provide evidence for the transport of anthropogenic PM from Asia on
Mt. Batchelor, OR by Jaffe et al. (2003). These data show sporadic but well correlated increases in CO,
03, total Hg, and aerosol backscatter associated with air coming from Asia. The ITCT-2K2 campaign also
found evidence for the oxidation of S02 to H2S04 during trans-Pacific transport of Asian emissions. If
particulate S042 were to be formed in the polluted boundary layer where it originated, it would likely be
deposited prior to transport across the Pacific Ocean (Brock et al., 2004). Thus primary species emitted
directly and secondary species formed during transport contribute to PRB levels. Satellite data have
provided images to track clouds of dust and pollution across the oceans. They have been used for some
quantitative estimation of the flux of material leaving continents. Yu et al. (2008) used optical thickness
data to estimate column loadings from the Mresolution Imaging Spectrometer (MODIS) along with
satellite assimilated wind fields to estimate the transport of PM from Asia. Three-dimensional,
global-scale, chemistry-transport models have also been used to estimate intercontinental transport of PM
pollution (UNCEC, 2007) and trans-Pacific transport of mineral dust from Asian deserts (Duncan Fairlie
et al., 2007) and the Sahara Desert (McKendry et al., 2007).
3.6.1.1. Estimating PRB Concentrations
Estimates of the distribution of daily average PRB concentrations in the 2004 PM AQCD
(U.S. EPA, 2004) were based on data obtained at IMPROVE monitoring sites in the West. Western sites
were chosen because they were thought to be among the least likely influenced by regional pollution
sources especially at the upper end of the concentration distribution. This conclusion was drawn from
back trajectory analyses and examination of the trace elemental composition. Estimates of PRB levels
December 2008
3-116
DRAFT-DO NOT QUOTE OR CITE
-------
1 reported in prior AQCDs for PM (U.S. EPA, 1996, 2004) were based in large measure on estimates by
2 Trijonis (1990) for the NAPAP as shown in Table 3-21.
Table 3-21. Estimates of annual average natural background concentrations of PM in different
size fractions (|jg/m3) from previous reviews.
PM2.5 PM10
PM10-2.5
East 2-5 5-11
<1-9
West 1-4 4-8 <1-7
Vouaqeurs
Figure 3-60. IMPROVE monitoring site locations.
3 Table 3-22 shows annual and quarterly average PM2 5 measurements from IMPROVE sites shown
4 in Figure 3-60 for 2004. Annual average concentrations tend to be higher in the East than in the Midwest
5 or West. However, when the data are broken down by season, more variability is notable. Highest values
6 in the East are found during the 3rd calendar quarter, whereas in the West highest quarterly averages can
7 occur during other quarters. As can also be seen from a comparison with values shown in Table 3-21,
8 values measured in the East are much higher than the estimated PRB levels.
December 2008
3-117
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Table 3-22. Annual and quarterly mean PM2.5 concentrations (ng/m3) measured at IMPROVE sites
in 2004.
Mean
January-March
April-June
July-September
October-December
EAST
Acadia
4.5
3.9
4.6
6.0
3.5
Brigantine
9.5
8.1
11.3
11.6
7.3
Dolly Sods
9.5
6.7
9.8
15.5
5.7
MIDWEST
Voyageurs
3.8
4.1
3.1
4.2
3.6
WEST
Bridger
2.1
1.2
3.1
2.8
1.3
Canyonlands
2.6
2.2
3.2
2.9
2.1
Gila
2.9
2.0
4.0
3.8
1.8
Glacier
4.8
4.6
4.2
5.3
5.0
Redwood
3.5
2.7
3.6
3.7
3.9
Thus, estimating daily average PRB levels in the eastern U.S. using observations
is highly
problematic because of the widespread mixing of precursors and anthropogenic PM generated in the East.
Two approaches were considered in the last NAAQS review. The first was to use the results of receptor
modeling studies to separate contributions from likely regional pollution sources from natural and
imported pollution. The second was to separate out components mainly thought to be emitted by regional
pollution sources such as S042 , which are obtained directly from observations at IMPROVE sites, and to
use the remaining PM components. Both of these approaches have limitations because receptor models
estimate the concentration profile but may suffer from inaccuracy of the assumptions used or limitations
in the grid approximations. Removal of regional pollutant sources involves assumptions about the
behavior of regional pollution that can lead to underestimation if all sources are not predicted.
3.6.1.2. CTM for Predicting PRB Concentrations
CTMs can be used to estimate the PRB concentrations of atmospheric components including PM
using a "zero-out" approach in which anthropogenic emissions inside the U.S., Canada, and Mexico are
set to zero and biogenic emissions remain, and both anthropogenic and biogenic emissions elsewhere in
the world remain. Numerical modeling can provide more precision in the estimate of PM PRB than
measurements since even the most remote measurement sites like some of those in the IMPROVE
network (see the discussion in Section 6.1.1. above) will necessarily be affected by non-local non-
biogenic pollution, thereby confusing the contributions from these sources.
December 2008
3-118
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
For this assessment, we have coupled the global-scale circulation model GEOS-Chem with the
regional scale air quality model CMAQ (see the discussion of CMAQ in Section 3.5.4.) to simulate one
year of air quality data over the CONUS in two series of runs, the first with all anthropogenic and
biogenic emissions included and the second annual series with the zero-out approach described just
above. The models were set up in this way: GEOS-Chem version 7, with modifications to include
aromatic and biogenic secondary organic aerosol formation; emissions computed from a variety of
sources including the Global Emissions Inventory Activity (GEIA; Benkovitz et al., 1996), and Emissions
Database for Global Atmospheric Research, version 2 (EDGAR; Olivier et al., 1996; Olivier et al., 1999).
Particularized emissions in specific areas used the European Monitoring and Evaluation Program (EMEP;
(Auvray and Bey, 2005), BRAVO (Kuhns and Knipping, 2005) for Mexico, Streets et al. (2006) for Asia,
Martin et al. (2002) for additional NOx emissions from biofuels, lightning, and ship traffic, Bond et al.
(2004) for global primary organic aerosols, Cooke et al., (1999) and Park et al. (2003) for U.S. primary
organic aerosols. Biomass burning emissions are not climatological but were computed with GFEDv2
(Giglio et al., 2006; van der Werf et al., 2006) monthly values using active fire observations from
MODIS; global dust fields were computed off-line using GOCART (see emissions from DEAD
(http://dust.ess.uci.edu/dead/) to make annual adjustments to photolysis rates and heterogeneous-phase
chemistry.
The regional CTM was configured in this way: CMAQ version 4.7 (excluding the dynamic coarse
mode updates), using the SAPRC 99 chemical mechanism and AER05 aerosol module; emissions
processed through SMOKE (http://smoke-model.org) version 2.4 based on the 2004 projections from the
NEI with specific CEM, biogenics, and fire updates; MM5 version 3.7.4 with the Asymmetric Convective
Mixing, version 2.2, PBL scheme; and data nudging to analyze fields for winds and temperature.
Model Evaluation
Details from evaluations of the performance of a number of CMAQ applications are given in
Arnold et al. (2003), Eder and Yu (2006), Appel et al. (2005), and Fuentes and Raftery (2005).
In an annual simulation series for 2002 using CMAQ v4.6.1 in two 12 km domains for the CONUS
(see Figure 3-61), predicted concentrations of summertime pS04, often a major determinant of surface-
layer PM concentrations, were we 11-predicted by CMAQ at 12 km grid spacings, to within a factor of 2 at
nearly every point of comparison and with R2 > 0.8 across all three national networks (CASTNet,
IMPROVE and CSN); a more detailed description is included in the 2008 NOx-SOx ISA (U.S. EPA,
2008e). This result for CMAQ v4.6.1 for 2002 tracks the generally well-predicted S042 concentrations
found in most earlier CMAQ evaluations: see Mebust et al. (2003), Eder and Yu (2006), and Tesche et al.
(2006). Since pS042 concentrations are strongly a function of precipitation, care must be taken to ensure
December 2008
3-119
DRAFT-DO NOT QUOTE OR CITE
-------
1 that the meteorological solution driving individual CMAQ chemical applications produces precipitation
2 fields with low bias as discussed by Appel et al. (2008).
2DGBK_ml2v33_l ThmF MM lor Juris to Auguul 3DS3
WPflDVE (2002ae maGwM 12hmE| a
^ STN.;20CEa:_rw2vM_iarT€^:. a* /
CASTN«r ?UiiiEl /
M*a &
o _
§
6
0
2
12
8
14
4
Observation
Figure 3-61. 12-km EUS Summer S042- PM each data point represents a paired monthly averaged
(June/July/August) observation and CMAQ prediction at a particular IMPROVE, STN,
and CASTNet site. Solid lines indicate the factor of 2 around the 1:1 line shown
between them.
3 Wintertime pN03 (Figure 3-62) and total N03 (HN03 + pN03) (Figure 3-63) concentrations are
4 predicted as well by CMAQ; but N03 is a pervasively difficult species to measure and model. Still, at the
5 CASTNet nodes where the total N03 concentrations are higher than they are at all but a few of the
6 remote IMPROVE sites, CMAQ predicts concentrations for nearly every node to within a factor of 2 and
7 with an R2 >0.8.
December 2008
3-120
DRAFT-DO NOT QUOTE OR CITE
-------
TDDZicjiwl A311 JkmE NCD Tor -iHcambf la February 2QQ7
Marffty Average-
aWQ9 j u^'ffiS)
* 3tMy ^ ~ ¦*¦
* V> fcr- **#l.
N
0
2
6
a
4
Observation
Figure 3-62 12-krn EUS Winter nitrate PM, each data point represents a paired monthly averaged
(December/January/February) observation and CMAQ prediction at a particular
IMPROVE and STN site. Solid lines indicate the factor of 2 around the 1 1 line shown
between them.
2U0Z*c_mrBv33_12kmE TN03 lor Uecemlwr to Obrunry 200Z
c GASTNei '20Ce^;jT.9l?v33_1 ^.kF)
MariJ-ly Avenge
TKOS (i»g.'1ri3)
a
0
1
2
3
4
5
E
Qb.5crva.1kin
Figure 3-63. 12-km EUS Winter total nitrate (HNO3 + total pNOs), each data point represents a
paired monthly averaged (December/January/February) observation and CMAQ
prediction at a particular CASTNet site. Solid lines indicate the factor of 2 around the
1:1 line shown between them.
1 A "base case" in which conditions for 2004 including all the anthropogenic and natural sources both
2 within and outside of the U.S., Canada and Mexico was ran for comparison with measurements. A PRB
December 2008
3-121
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
simulation was also run by shutting off the anthropogenic sources of primary PM and precursors to
secondary PM in the U.S., Canada and Mexico.
Acadia
Brigantine
Monitored
Monitored
Voyageurs
Monitored
Monitored
Figure 3-64. Monthly average of PM2.5 concentrations measured at IMPROVE sites in the East and
Midwest for 2004. Also shown are distributions of PM2.5 concentrations calculated by
CMAQ for the base case and for PRB.
Figures 3-61 through 3-63 show monthly average concentrations, and Figures 3-64 through 3-66
show 24-h average concentration distributions for 2004 predicted by CMAQ for the base case and for
PRB and measurements at the IMPROVE sites shown in Figure 3-60. As can be seen from Figures 3-61
and 3-64 , CMAQ predicted base case concentrations are generally within 1 or 2 |ag/irf across the entire
concentration distribution at Eastern and Midwestern sites. There is an indication that wild fires affected
the grid cell containing the Voyageurs site, but that the site itself was not affected. The "base case"
simulations tend to underestimate concentrations throughout the concentrations at most western sites as
shown in Figures 3-62 and 3-65. These underestimates are still within the range of a few (.ig/ni \ However,
the base case simulation also greatly over-predicts PM2 5 concentrations at the upper end of the
distribution at the Redwoods site (Figure 3-66). This over-prediction results from emissions from wild
fires in northern California that are included in the grid cell containing the Redwoods site, but may not
December 2008
3-122
DRAFT-DO NOT QUOTE OR CITE
-------
1 have affected the site. However, wild fires indicated by MODIS would have affected other areas either
2 close to these sites or could have affected other locations in-between the IMPROVE sites.
Bridger
Canyonlands
Monitorec
Monitored
Monitored
Monitored
Figure 3-65. Monthly average of PM2.5 concentrations measured at IMPROVE sites in the West for
2004. Also shown are distributions of PM2.5 concentrations calculated by CMAQ for
the base case and for PRB. Note the scale change from the preceding figures.
Redwoods
12
10
8
6
4
2
0
2
3
4
5
6
7
8
9
10
12
Month
Figure 3-66. Monthly average of PM2.5 concentrations measured at the Redwoods National Park
IMPROVE sites in California for 2004. Also shown are distributions of PM2.5
concentrations calculated by CMAQ for the base case and for PRB. Note the scale
change from the preceding figures.
December 2008
3-123
DRAFT-DO NOT QUOTE OR CITE
-------
Acadia
min 10 20 30 40 50 60 70
Percentile
Monitored —¦— Base
90 95 99 max
PRB
Brigantine
min 10 20 30 40 50 60 70 80 90 95 99 max
Percentile
Monitored —¦—Base —±—PRB
Dolly Sods
Voyageurs
min 10 20 30 40 50 60 70
Percentile
Monitored —¦—Base
90 95 99 max
PRB
min 10 20 30 40 50 60 70
Percentile
90 95 99 max
Monitored -
PRB
Figure 3-67. Distribution of PM2.5 concentrations measured at IMPROVE sites in the East and
Midwest for 2004. Also shown are distributions of PM2.5 concentrations calculated by
CMAQ for the base case and for PRB.
December 2008
3-124
DRAFT-DO NOT QUOTE OR CITE
-------
Bridger
Canyonlands
Percentile
Percentile
Monitored
PRB
Monitored
PRB
Gila
Glacier
20
18
16
14
12
8
Q_
6
4
2
0
min 10 20 30 40 50 60 70 80 90 95 99 max
Percentile
Percentile
Monitored
PRB
Monitored
PRB
Figure 3-68. Distribution of PM2.5 concentrations measured at IMPROVE sites in the West for 2004.
Also shown are distributions of PM2.5 concentrations calculated by CMAQ for the base
case and for PRB. Note the scale change from the preceding figures.
December 2008
3-125
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
Redwoods
80
70
60
50
40
30
20
10
0
min 10 20 30 40 50 60 70 80 90 95 99 max
Percentile
Monitored —¦—CMAQ Est PRB
Figure 3-69. Distribution of PM2.5 concentrations measured at the Redwoods National Park
IMPROVE sites in California for 2004. Also shown are distributions of PM2.5
concentrations calculated by CMAQ for the base case and for PRB. Note the scale
change from the preceding figures.
Table 3-23 gives the annual and quarterly average CMAQ predictions at IMPROVE sites for the
"base case" and the ratio of CMAQ predictions to the measured concentrations at those sites in 2004.
CMAQ performance for the annual average concentrations and most of the seasonal averages is very
good in the East and Midwest, generally falling within 35%. In the West, CMAQ's prediction of PM2.5
mass averages at these remote sites is generally too low in all seasons, often by 50%. Air quality model
predictions in the mountainous West are often less-good than over the flatter terrain in the East and
Midwest because the model's grid spacing (36 km in this case) smoothes over significant variation at the
surface which results in differences at such remote sites. However, the model's trend relative to the
geospatial difference is correct: the predicted PM2.5 concentrations are lower at the western sites than
they are in the East, just as the measurements are. Table 3-24 shows corresponding PRB PM2 5
concentrations at IMPROVE sites.
December 2008
3-126
DRAFT-DO NOT QUOTE OR CITE
-------
Table 3-23.
Annual and quarterly mean PM2.5 concentrations (ng/m3) for the CMAQ "base case" at
IMPROVE sites in 2004.
Annual; mod/obs
Jan-March; mod/obs
Apr-Jun; mod/obs
Jul-Sep; mod/obs Oct-Dec; mod/obs
EAST
Acadia
4.7; 1.04
5.6; 1.44
4.0; 0.87
4.6; 0.77
4.6; 1.31
Brigantine
10.2; 1.07
10.9; 1.35
10.3; 0.91
10.2; 0.88
9.4;1.29
Dolly Sods
9.8; 1.03
8.3; 1.24
8.6; 0.88
14.0; 0.90
8.0; 1.40
MIDWEST
Voyageurs
4.0; 1.05
4.9; 1.19
2.2; 0.71
3.9; 0.93
4.9; 1.36
WEST
Bridger
1.6; 0.76
1.3; 1.08
1.6; 0.52
1.8; 0.64
1.7; 1.30
Canyonlands
1.6; 0.62
1.9; 0.86
1.4;0.44
1.5; 0.52
1.6; .76
Gila
1.6; 0.55
1.4; 0.70
2.2; 0.55
1.7;0.45
1.1; 0.61
Glacier
2.2; 0.45
1.8; 0.39
2.1; 0.50
2.1; 0.40
2.8; 0.56
Redwood
4.6; 1.31
4.0; 1.48
3.0; 0.83
2.9; 0.78
8.4; 2.15
Table 3-24.
Annual and quarterly mean PM2.5 concentrations (ng/m3) for the CMAQ PRB
simulations at IMPROVE sites in 2004.
Annual
Jan-March
Apr-Jun
Jul-Sep
Oct-Dec
EAST
Acadia
0.70
0.76
0.76
0.65
0.65
Brigantine
0.77
0.86
0.91
0.70
0.63
Dolly Sods
0.79
0.88
0.83
0.75
0.66
MIDWEST
Voyageurs
1.2
0.83
0.91
2.0
0.93
WEST
Bridger
0.57
0.33
0.57
0.76
0.61
Canyonlands
0.49
0.38
0.54
0.68
0.35
Gila
0.74
0.42
1.4
0.80
0.32
Glacier
0.91
0.36
0.87
1.1
1.3
Redwood
2.8
2.4
1.5
1.1
6.1
1 Table 3-25 illustrates CMAQ predictions of seasonal variation in the base case PM2 5
2 concentrations across regions of the CONUS, while Table 3-26 shows CMAQ predictions of the seasonal
3 variation in regional PRB PM2 5 concentrations. Highest base case PM2 5 concentrations were observed for
December 2008
3-127
DRAFT-DO NOT QUOTE OR CITE
-------
1 the Northeast, Southeast, and Industrial Midwest with highest concentrations during the fall and winter
2 (and comparably high concentrations in the summer for the Industrial Midwest), while PRB PM2 5
3 concentrations were highest annually in the Southeast, which has highest levels during the winter. In the
4 summer, PRB PM2 5 is comparable for the Northwest and Southeast and elevated but slightly lower for the
5 Southwest. These observations also likely suggest the influence of wildfires in the west.
Table 3-25.
Annual and quarterly mean of the CMAQ-predicted base case PM2.5 concentrations
(|jg/m3) in the U.S. EPA CONUS regions in 2004.
Annual
January-March
April-June July-September October-December
Northeast
9.76
10.74
8.38 9.55
10.38
Southeast
10.05
12.28
7.72 9.78
10.42
Industrial Midwest
11.38
12.22
9.37 11.89
12.00
Upper Midwest
6.70
8.83
4.95 5.34
7.67
Southwest
3.30
4.08
2.77 3.31
3.03
Northwest
2.72
2.49
2.21 2.71
3.44
Southern California
4.43
4.64
3.93 4.34
4.82
Table 3-26.
Annual and quarterly mean of the CMAQ-predicted PRB PM2.5 concentrations (ng/m3)
in the U.S. EPA CONUS regions in 2004.
Annual January-March April-June July-September
October-December
Northeast
0.74
0.85 0.78 0.67
0.68
Southeast
1.72
2.43 1.41 1.41
1.64
Industrial Midwest
0.86
0.89 0.89 0.94
0.73
Upper Midwest
0.84
0.79 0.93 0.99
0.66
Southwest
0.62
0.61 0.76 0.70
0.40
Northwest
1.01
0.48 0.81 1.42
1.32
Southern California
0.84
0.54 0.92 1.21
0.67
December 2008
3-128
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
3.7. Issues in Exposure Assessment for PM and its
Components
3.7.1. Introduction and Key Concepts
The purpose of this section is to present the latest exposure assessment studies for the purpose of
aiding the interpretation of epidemiologic studies described in subsequent chapters of this Integrated
Science Assessment. A theoretical model of personal exposure is presented to highlight what is
measurable and what uncertainties exist in this framework.
An individual's daily exposure to airborne PM can be described based on a compartmentalization
of the person's activities throughout a day:
E = |C jdt
Equation 3-2
where E = 24-h exposure, Q = airborne PM concentration at location /. and dt = portion of the day spent
in location j. This basic equation was broken down by Wu et al. (2005) into a microenvironmental model
that accounts for exposure to pollutants generated indoors and outdoors of the form:
E = f0C0+JdfiCi
Equation 3-3
where/= fraction of a day, subscript o = outdoor, and subscript i = indoor. Note that f, + I/,' = 1, and the
indoor term has a summation because it reflects various microenvironmental exposures. Here, "indoors"
refers to being inside any aspect of the built environment, e.g. home, office buildings, enclosed vehicles
(automobiles, trains, buses), and/or recreational facilities (movies, restaurants, bars). "Outdoor" exposure
can occur in parks, yards, sidewalks, and on bicycles or motorcycles. It is proportional to the ambient
concentration detected at monitoring sites. The complex human activity patterns across the population (all
ages) are illustrated in Figure 3-70 (Klepeis et al., 2001). This figure illustrates the diversity of daily
activities among the population as well as the proportion of time spent in each microenvironment.
December 2008
3-129
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
100
90
SO
Office/Factor}''
Near Vehicle
(Outdoors)
Mall/Store
Inside Vehicle
Residence-Outdoors
Residence-Indoors
OlllCf OutdOOI'
Other
Indoor
Bar,'Restaurant
School/
Dublin TD\Ar*
r Liu
qqj=jsssssKj=jKPiKsq
TO ft ft ft ft ft ft ft ft A ft ft ft cd
ooooooooooooooooooooooooo
OOOOOOOOOOOOOOOOOOOOOOOOO
ri ^ r-i r'n Ti ii h (X) oi o ^ ri fH r-i rn ^ tI i h 00 i o -h
Source: Klepeis et al. (2001)
Figure 3-70. Distribution of time sample population spends in various environments, from the
National Human Activity Pattern Survey.
In general, the relationship between personal exposures and ambient concentrations is complex
because ambient pollutants can be lost through chemical and physical processes in microenvironments
and the composition of PM can be modified during infiltration of outdoor air into microenvironments
(Meng et al., 2007a; Sarnat et al., 2006b).
The infiltrated outdoor PM is a function of outdoor PM and infiltration factors specific to the
ventilation properties of the building or vehicle in which the person is exposed:
a = F-mf c„
Equation 3-4
mf 0
where Finf = Pa/(ci + k) = infiltration factor (assuming the indoor mass balance is at steady state),
P = penetration coefficient, a = air exchange rate and k = particle loss rate. Finfquantifies the equilibrium
fraction of the PM concentration outside the microenvironment that penetrates inside the
microenvironment and remains suspended. It is a function of the building air exchange characteristics and
the particle properties. Properties of Fmf and studies of infiltration are summarized in Section 3.7.3.
If there are no significant local, outdoor sources and sinks of PM, then Ca, the concentration
measured by an ambient monitor can be approximated by C0. In this case, the ambient component of a
pollutant's microenvironmental concentration can be represented as the product of the ambient
December 2008
3-130
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
concentration and the infiltration factor (Fmf or a [if people spend 100% of their time indoors]). Alpha, a,
is the ratio of personal exposure from a pollutant of ambient origin to the pollutant's ambient
concentration (or the ambient exposure factor); a varies between 0 and 1. If local sources and sinks exist
and are significant, then the ambient component of outdoor air must be estimated using dispersion
models, land use regression models, receptor models, fine scale chemistry-transport models or some
combination of these techniques (cf. Section 3.7.2.3).
A variety of approaches can be used to estimate exposure to ambient PM. In some cases, individual
personal exposures are measured with personal exposure monitors (PEMs), where personal samples are
taken to estimate population exposure. In other cases, ambient concentrations are used as an exposure
surrogate. The ambient concentration may be based on measurements made at a single ambient monitor,
or as the average of several ambient monitors. If appropriate measurements are made, it is also possible to
estimate the ambient and non-ambient components of personal exposure. Two studies have used average
relationships for a panel to infer average relationships for a larger cohort (Dominici et al., 2000; Strand et
al., 2006), one study has used the average relationships for each member of a panel (Koenig et al., 2005),
and one study has estimated the unmeasured ambient and non-ambient components of personal exposure
for each panelist (Ebelt et al., 2005; Wilson and Brauer, 2006). Better associations are obtained between
health effects and ambient exposure than between health effects and total personal exposure. Wilson and
Brauer's (2006) results indicate that exposure error is introduced by: (1) using ambient concentrations
instead of ambient exposure (ambient concentrations showed an association with the effect but it was not
statistically significant): and (2) assuming ambient exposure and non-ambient exposure have the same
effects on health outcomes, i.e., identical toxicity, (there was essentially no association of the effect with
total personal exposure or the non-ambient component).
The use of personal exposure in population exposure assessment studies could cause various errors
in the health effect estimate because ambient and non-ambient particles differ in size distribution and
chemical composition (Ebelt et al., 2005; Long et al., 2001; Wilson and Suh, 1997), and the correlation
between personal exposure and ambient concentration may be different for each subject and may not be
statistically significant (Wilson et al., 2007b). Moreover, the use of ambient concentration could cause
error because the relationship between ambient concentration and personal exposure may be different for
each subject in the panel. The correlation between ambient concentration and personal exposure may be
high for some subjects, in which case the exposure error caused by using ambient concentration instead of
personal exposure may be small. In other subjects, the correlation may be low or negative (and not
statistically significant). In this case, the exposure error will be high and may obscure relationships
between ambient exposure and health effects. Table 3-27 shows the correlations between ambient
concentration and personal exposure and the t-statistic for the one panel (Ebelt et al., 2005; Wilson et al.,
December 2008
3-131
DRAFT-DO NOT QUOTE OR CITE
-------
1 2007b) (the relationship was statistically significant only for 5 of the 16 subjects). However, the
2 relationships between ambient concentration and the ambient component of personal exposure were
3 statistically significant for all subjects. These differences in correlation occur in part because the ambient
4 exposure factor, a, is different for each subject as shown in Table 3-27. Therefore, ambient concentration
5 will be better than total personal exposure as a surrogate for ambient exposure.
Table 3-27. Statistical parameters for the relationships between exposures and ambient
concentrations for each subject separately (values of C based on average of five
monitoring sites, E and A outliers included) sorted according to the correlation
coefficient of E with C.
Subject
1RforE
2t for E
3rforA
4a
13
0.83
3.35
0.92
0.71
2
0.82
3.15
0.95
0.78
4
0.74
2.47
0.75
1.04
15
0.74
2.17
0.89
0.82
1
0.72
2.33
0.90
0.64
11
0.68
1.85
0.99
0.96
3
0.66
1.98
0.96
0.83
5
0.51
1.20
0.86
0.70
8
0.44
1.09
0.87
0.95
14
0.40
0.96
0.56
0.39
7
0.20
0.45
0.87
0.72
10
0.08
0.18
0.90
0.54
12
0.02
0.04
0.98
0.64
9
-0.19
-0.42
0.97
0.94
6
-0.28
-0.65
0.90
0.73
16
-0.68
-1.59
0.77
0.62
Average
0.36
0.88
0.75
STD
0.45
0.11
0.18
CV
1.26
0.13
0.24
1 Pearson's correlation coefficient, R, for the correlation of the specified variable with the corresponding ambient concentration.
2The t-statistic for the slope of the regression of E against C (t >2 indicates statistical significance at the p <0.05 level).
3AII values of r are statistically significant.
4Average, for each subject separately, of individual values of a ij estimated using aij = Eij/Ci for SO42" data as an estimate of aij.
Source: Wilson and Brauer(2006)
December 2008
3-132
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
3.7.2. Methods for Estimating PM Exposures
The purpose of this section is to present new discoveries related to measuring and modeling aerosol
concentration and the error involved in these endeavors. A review of over 200 personal and
microenvironmental PM exposure studies published since 2002 (see Annex A) reveals that the majority of
the monitoring and modeling techniques in use were previously reviewed in the 2004 PM AQCD
(U.S. EPA, 2004) for PM. Detailed descriptions of these methodologies are provided in the 2004 PM
AQCD and therefore will not be repeated in this document. The following sections will include only
findings from 2002 or later regarding monitoring and modeling methodologies in common use and
significant advancements in understanding the capabilities and limitations of these methods for
assessment of PM exposure.
3.7.2.1. Exposure Monitoring and Associated Instrumental Measurement Errors
New Developments in Personal Exposure Monitoring Techniques
Personal exposure sampling methodology consists largely of integrated filter sampling using a
cyclone or Personal Exposure Monitor (PEM) to achieve a cut-point at a desired particle size (e.g., Hopke
et al., 2003; Larson et al., 2004). This method of sampling facilitates speciation work because the filters
can be archived for chemical and gravimetric analysis. Additionally, light scattering aerosol detection
instruments, such as the personal DataRam (pDR) and SidePak personal aerosol monitor have seen some
usage in personal PM10, PM2 5, and PM, monitoring (e.g., Lewne et al., 2006). However, because these
technologies are not new, the reader is again referred to the 2004 PM AQCD for further information on
these techniques.
One area of further development is in personal sampling of the thoracic and respirable particle size
distribution. Variations of the cascade impactor have been developed for personal sampling and tested for
use in field studies (Case et al., 2008; Lee et al., 2006a; Singh et al., 2003). The models developed and/or
tested by (Lee et al., 2006a) and Case et al. (2008) operate with a 1 ^m cut point and therefore can
characterize respirable particles well. The Personal Cascade Impactor Sampler (PCIS) has the capability
to sample down to a cutpoint of 250 nm (Singh et al., 2003). For PM2 5, the difference between the PCIS
and MOUDI cascade impactor was 11%, while difference between the PCIS and SMPS-APS was only
2%. Difference in PM25 species compared with the MOUDI was generally higher: 11% for S042 , 22%
for N03~, 19% for EC, and 94% for OC. Mass was overestimated by 3%, 16%, and 31% for PMi_0.5,
PMo.5_o.25, and PM0.25, respectively, when compared with the SMPS-APS. Similarly, Case et al. (2008)
found a difference ranging from -11 to +10% for PM10-2.5 with the Personal Respirable Particulate
Sampler (PRPS), and Lee et al. (2006a) found a difference of -6 to 0% for PM2 5 and -6 to -1% for PMi0
December 2008
3-133
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
when comparing this device with the PEM. Leith et al. (2007) also tested a passive PM sampler for
detection of PM10.2.5 and found the difference between a PM10_2.5 FRM and the co-located passive sampler
was within lo of concentrations measured by the FRM samplers.
New Developments in Microenvironmental Exposure Monitoring Techniques
The majority of developments since the 2004 PM AQCD regarding microenvironmental PM
characterization have involved real-time instrumentation in the ultrafine PM size range. Because these
methods are also used for ambient sampling, they are described in section 3.4.
New developments in microenvironmental sampling for exposure assessment have also included
construction, testing, and implementation of mobile environmental sampling laboratories for PM mass,
particle count, and composition, as well as other criteria pollutants (CO, SOx, NOx, 03). These mobile
laboratories typically contain a suite of real-time equipment with short sampling intervals (e.g. 1-10
minutes), such as an SMPS with CPC, APS, laser photometers, and aethalometers for aerosols; FRMs for
the gaseous criteria pollutants; weather station for meteorological variables, and a Global Positioning
System (GPS) for position. Videotape or journal observations are sometimes logged simultaneously to
track local on-road sources of pollution in this way. One key application of mobile laboratory technology
is assessment of the outdoor microscale environments and in-vehicle microenvironments on roadways for
determining exposure during on-road transportation (Piijola et al., 2004; Sabin et al., 2005; Weijers et al.,
2004; Westerdahl et al., 2005). For instance, Sabin et al. (2005) used videotape records to determine
whether BC detected on a school bus was the result of local outdoor sources from other vehicles or
"self-pollution" from the school bus's own engine exhaust. Westerdahl et al. (2005) used the GPS time
series to determine that minima in the ultrafine PM time series corresponded to passage through
residential areas of Long Beach and Pasadena, in contrast to the pollution spikes observed along
highways. Studies have also shown that detection of automobile and DE could be improved through use
of combined measurement results to improve statistical analysis (Ntziachristos and Samaras, 2006) and
identification of best tools for measurement (Kinsey et al., 2006).
Measurement Error at Community-Based Ambient Monitors and Exposure
Assessment
Section 3.4 discusses potential errors in measuring ambient PM in detail. Because there will likely
be some random component to instrumental measurement error, the correlation of the measured PM mass
with the true PM mass will likely be less than 1. Sheppard et al. (2005) indicate that instrument error in
the individual or daily average concentrations have "the effect of attenuating the estimate of a." However,
Zeger et al. (2000) state that the "instrument error in the ambient levels is close to the Berkson type" and
in order for this error to cause substantial bias in later estimation of the health outcome, the error term (the
December 2008
3-134
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
difference between the true concentrations and the measured concentrations) must be strongly correlated
with the measured concentrations. Zeger et al. (2000) suggest that, "Further investigations of this
correlation in cities with many monitors are warranted."
Measurement Error for Personal Exposure Monitors
PEMs are specialized monitors that, because people must carry them, have to be small, light, guiet,
and battery operated or passive. As a result, they may have lower flow rates and pressure drops for filter
measurements of PM or light scattering measurements that are not sensitive to ultrafine or smaller
accumulation mode particles. Thus, there may be considerable differences between the ambient
measurement and the PEM measurement of the same indicator. This is especially a problem for
semi-volatile components such as OC and NH4NO3. These errors are described in much greater detail in
the 2004 PM AQCD (U.S. EPA, 2004).
Exposure Error Associated with Community-Based Ambient Monitor Height
Community-based ambient monitor height can affect estimates of exposure. Maletto et al. (2003)
measured vertical accumulation mode, fine, and coarse particle distributions for elevations up to 1000 m
with a suite of instruments and demonstrated substantial variability in the vertical profile. Their results
show significantly higher concentrations at ground level in the fine and coarse mode and substantially less
variation in the accumulation mode during daytime. In the case of community-based ambient monitoring,
variation closer to ground level is more relevant because monitors are not sited more than 14 m above
ground level (Watson et al., 1997). Wu et al. (2002) performed outdoor monitoring at heights of 2, 8, 19,
30, 59, and 79 m above ground level along a street with average traffic of 1485 vehicles per hour. When
moving from 2 m to 8 m above ground level, average measured concentration dropped by 26% for PMi0,
12% for PM2 5, and 15% for PM,. Over the measurement heights from 2 m to 79 m, averaged measured
concentration dropped by 40% for PM10, 37% for PM2 5, and 19% for PM,. These findings suggest that
measurements from elevated community ambient monitors underestimate ground-level outdoor
exposures.
3.7.2.2. Uncertainties in PM Exposure Assessment
Relationship between Community-Based Ambient Measurements and Personal
Exposure
Intra-urban spatiotemporal variability (i.e. central site concentrations vs. concentrations outside a
subject's home), as well as infiltration properties related to particle size and composition, are important
factors in the relationship between ambient measurements and personal exposures. Filleul et al. (2005)
December 2008
3-135
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
computed exposure based on varying contributions of community-based ambient monitors and proximal
monitors (to represent a receptor) and found that increasing the weight of proximal monitors resulted in
non-significant but increased mean concentrations. Meng et al. (2004) examined representation of the
ambient contribution to personal exposure using four models with increasing detail regarding differential
infiltration and residential ventilation properties. The authors found that both the contribution of ambient
PM to indoor exposures and the variability in those estimates increased with increased model detail. In
other studies comparing ambient concentrations with indoor concentrations and personal exposures, Liu
et al. (2003) and Williams et al. (2008) both noted large differences between central site monitor data and
data obtained from personal and indoor PMi0 and PM2 5 samplers. Williams et al. also noted this for
PM10-2 5, as shown in Figure 3-71.
Spatial variability among various studies further suggests that use of a single or small number of
ambient monitors introduces uncertainty in exposure assessment studies. Violante et al. (2006) studied
personal exposures to traffic of parking police in Bologna, Italy to determine how personal exposure to
outdoor PM10 and benzene compares with that measured at a community-based monitor. This study found
that personal exposures to PMi0 were significantly higher than at the community-based monitor, although
the authors were not able to demonstrate significant effects of meteorology or traffic on those exposures.
Spatial heterogeneity of personal exposures to metals in PM10 and PM2 5, with higher levels found near
high-traffic and industrial areas, was observed by Nerriere et al. (2007). In a Bayesian hierarchical model
analysis of personal exposure and ambient data from the pilot Baltimore Epidemiology-Exposure Panel
Study of 16 subjects, McBride et al. (2007) showed that community monitors overestimated personal
exposures, and that these results were not sensitive to model selection. Likewise, significant differences
among personal, indoor, and outdoor community monitors in studies by Adgate et al. (2002; 2003) and
Kim et al. (2005b) were also reported. The results of these studies are described further in Section 3.7.3,
but they are important to mention here in the context of representativeness of community monitored
ambient PM across an urban population.
December 2008
3-136
DRAFT-DO NOT QUOTE OR CITE
-------
"e
u>
o
'E
o
S
S
Q_
E
3
E
E
o
o
Personal PM10.2.5 Monitor (^g m'"5)
Source: Williams (2008)
Figure 3-71. Comparison of community dichot and personal exposure monitor for PM10-2.5-
Reiationship between Community-Based Ambient Measurements and
Community Averaged Concentrations
Exposure Assessment for Community Time-Series Epidemiology
1 For community time-series epidemiology, the community averaged concentration, not the
2 concentration at each fixed monitoring site, is the concentration variable of concern (Zeger et al., 2000).
3 The correlation between the concentration at a central community ambient monitor and the community
4 averaged concentration depends on at least the following three factors:
5 ¦ Even distribution of the indicator across space: Regional pollutants such as S042 will be
6 more evenly distributed than point source pollutants. Traffic emissions might show spatial
7 heterogeneity near sources but more homogeneous distribution farther downwind from
8 sources.
Personal = 0.2789*Community + 10.282
R2 = 0.1169 *
December 2008
3-137
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
¦ Selection of the monitoring site chosen to represent the community average: If the site is
selected to measure a "hot spot" or pollution from a nearby source, estimates across the
community could be skewed upwards.
¦ Division of the community by terrain features or source locations into several
sub-communities that differ in the temporal pattern of pollution: Intra-urban spatial
heterogeneity is discussed in detail in Section 3.5. Community exposure may not be well-
represented when monitors cover large areas with several sub-communities having different
sources and topographies. This point is illustrated for Los Angeles in Figures 3-27 and 3-36.
In Phoenix, the use of ZIP-code classified mortality data enabled researchers to find high risk ratios
in a small area near the monitoring site (Mar et al., 2003; Wilson et al., 2007b) while use of county-wide
data produced non-significant associations (Moolgavkar, 2000; Smith et al., 2000). It seems likely that at
least part of the heterogeneity found between cities in multicity studies is due to the use of a geographic
area that is composed of several sub-communities that differ in the spatiotemporal distribution of air
pollutants. For all metropolitan areas investigated in this assessment, the PMi0 data have significantly
more scatter, which suggests that the uncertainty of the community average concentration would increase
in the coarse PM range. Metrics have been developed and used to compare the spatial variability of air
pollutants (Wongphatarakul et al., 1998). These metrics are useful in assessing the potential for exposure
error in the epidemiologic studies, especially when different monitors are used on different days to
construct city-wide averages.
Long-Term Exposure Assessment for Epidemiology Studies
Epidemiologic studies of long-term exposure rely on differences among communities in long-term
average ambient concentrations. If exposure errors are different in the different communities, the
differences in long-term ambient concentrations among communities may not represent the differences in
long-term average exposures. For example, there may be community to community differences in
measurement error, in the average ambient exposure factor (a) or the average non-ambient exposure. This
could happen if exposure to fresh pollutants generated by vehicular traffic or pollutants from other
localized sources differed among the spatial areas. Thus, in a regression of health effects against average
concentration (as an indicator for average exposure) there could be a different amount of error (either
positive or negative) in the exposure indicator for each spatial area. This could add error and bias the
slope up or down. There is a general progression toward the use of concentration fields that account for
spatial variations in concentration. However, it has not yet been possible to include individual or
small-area variations in the personal exposure to ambient concentrations or variations in personal
December 2008
3-138
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
exposures to indoor-generated pollutants in long-term studies of the associations of pollutants with health
effects.
3.7.2.3. PM Exposure Modeling
Exposure Modeling Techniques
Stochastic Population Exposure Models
Section 3.7.1 describes the conceptual model representing an individual's exposure to PM of
ambient and non-ambient origin. A number of techniques to describe these exposures at the individual and
population level have been published since 2002. Population-based methods, such as the Air Pollution
Exposure (APEX), Simulation of Human Exposure and Dose System (SHEDS), and EXPOLIS (exposure
in polis, or cities) models, involve stochastic treatment of the model input factors
(http ://www.epa.gov/ttn/fera/human apex.html (Burke et al., 2001; Kruize et al., 2003). Another approach
is to predict location-based exposures using a deterministic model such as the CMAQ model, California
Line Source Dispersion Model (CALINE), CALPUFF (long-range plume transport model), and
Operational Street Pollution Model (OSPM) for determination of street-level PM pollution coupled with
infiltration models to represent indoor exposure to ambient levels (Appel et al., 2008; Gilliam et al., 2005;
Hering, 2007; Mensink et al., 2008; Wilson and Zawar-Reza, 2006). Stochastic and deterministic methods
are often combined, as described below. Land use regression (LUR) models have also been developed to
describe pollution levels as a function of source behavior (Briggs et al., 1997; Gilliland et al., 2005; Ryan
and LeMasters, 2007). These are described in detail in Annex 3.7 of the 2008 NOx ISA (U.S. EPA,
2008c). Recent developments are described in this subsection, and relevant errors in these approaches are
described in the following sub-section.
Stochastic models, such as SHEDS, APEX, and EXPOLIS utilize a distribution of pollution and
individual-level variables, such as ambient and local PM concentration source contributions and breathing
rate respectively, to compute the probability of individual exposure. Recently, SHEDS has been linked
with the Modeling Environment for Total Risk Studies (MENTOR) model to expand population exposure
assessment to individual risk assessment (Georgopoulos et al., 2005). In this formulation, CMAQ was
used to predict initial concentrations at a coarse scale, and then a spatiotemporal random field method
(Vyas and Christakos, 1997) or Bayesian max entropy method (Serre and Christakos, 1999) was applied
to interpolate the concentration to census tract scale in which exposure estimates are made. Activity
diaries such as the Consolidated Human Activity Database (CHAD) (McCurdy et al., 2000) are employed
to estimate the time basis of microenvironmental exposures, and then specific microenvironmental
exposures are assessed by utilizing distributions of parameters such as air exchange rate. Finally, within
December 2008
3-139
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
MENTOR, the estimates of exposure are related to dose and metabolic distributions to estimate risk of
specific health impacts.
Dispersion Models
Dispersion models have been used both for direct estimation of exposure and as inputs for
stochastic modeling systems, as described above. For instance, CALPUFF was used to model transport
and dispersion in lower Manhattan following the September 11, 2001 World Trade Center collapse
(Gilliam et al., 2005) to determine average location-based exposures, and Wilson and Zawar-Reza (2006)
used the MM5 model to assess PMi0 dispersion and potential for exposure in Christchurch, New Zealand.
In a method similar to that employed by Georgopoulos et al. (2005) with SHEDS, Wu et al. (2005) used
CALINE to predict street-level concentrations of pollutants and input the results of that dispersion model
into an individual exposure model that accounts for infiltration of specific building characteristics. Wu et
al. also employed CHAD to estimate the time-basis of exposures. With an individualized exposure
approach, the model is deterministic. However, population exposures can be estimated by performing
repeated simulations using various housing characteristics and then computing a posterior probability
distribution function for exposure.
Land Use Regression Models
LUR models are also applied to individuals, primarily at the intra-urban scale (Ryan and
LeMasters, 2007). At the census tract level, an LUR is a multivariate description of pollution as a function
of traffic, land use, and topographic variables (Briggs et al., 1997). Originally, LUR was used for N02
dispersion, but it was adapted for PM2.5 exposure estimation by Brauer et al. (2003) for Stockholm,
Sweden, Munich, Germany, and throughout The Netherlands. This study found a measure of traffic
density to be the most significant variable predicting PM2 5 exposure. Ryan et al. (2008) reported on an
LUR model for childhood exposure to traffic-derived EC for the Cincinnati Allergy and Air Pollution
Study and also found traffic to be the most important determinant of pollution. In this case, wind direction
was also factored into the model as a determinant of EC mixing. Like deterministic dispersion models,
LUR can be performed over wide areas to develop a posterior probability distribution function of
exposure at the urban scale.
Source proximity is sometimes used as a surrogate for exposure. For instance, Baxter et al. (2007a)
predicted indoor exposure to PM2 5, EC, and NOx based on distance to traffic sources, indoor source
characteristics, window opening, and ambient concentrations in the Boston metropolitan area. In this
effort, Baxter et al. examined a variety of factors obtained from GIS including roadway density, roadway
length, average daily traffic, and population density to determine which variables were significant
December 2008
3-140
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
predictors. They found that PM2 5 was largely influenced by ambient PM2 5 while EC was more sensitive
to local traffic sources. In a different approach, Corburn (2007) tested two distinct metrics, the cumulative
air toxics surface (CATS) and the U.S. EPA's National Air Toxics Assessment (NATA) to determine how
these approaches can yield estimates of human exposure to diesel and 33 air toxics for environmental
impact assessment. The CATS approach included an exposure term incorporating source density and
distance to source, and the sources could include traffic as well as bus depots and transfer stations,
airports, and industrial point sources. Corburn's paper demonstrated that robust land use data can provide
an approximation for urban exposures, although he cautioned that such estimates should not supersede
environmental monitoring but be used in arguing for added monitoring. In using these approaches, Huang
and Batterman (2000) warn that geographic divisions must be sufficiently small to avoid inter-zone
variability in source and exposure characteristics.
Model-Related Errors
Model-related errors are determined by four factors: accuracy of the mathematical model, accuracy
of model inputs, scale model resolution, and model sensitivity. If verification errors related to these four
factors are minimized, then the model can beevaluated against physical data to determine how well the
model truly captures a real situation (Roache). Detail of the model design and inputs can have significant
impact on validation, as demonstrated in Meng et al. (2005b) and Hering et al. (2007). Meng et al.
demonstrated how use of an increasingly more detailed mathematical model decreases the variability of
the results with respect to modeled indoor PM2 5 concentration of outdoor origin and infiltration factor.
Hering et al. compared infiltration model results for PM-based EC, N03 . and S042 . Model inputs were
from a central site monitor only, central site monitor with air exchange data, and detailed inputs related to
initial outdoor (outside test building) and indoor concentrations. Use of more detailed inputs resulted in
significant reductions in error for indoor EC concentration, smaller improvements for indoor S042 . and
negligible improvement in model results for indoor N03 . Observed errors are consistent with
observations in the literature regarding adequacy of community based monitors for epidemiology studies
cited in Section 3.7.2.1 above. This distinction also illustrates the strong impact of differential infiltration
discussed in Section 3.7.3.
December 2008
3-141
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
4450
4440
o
Z
£
"3
o
w
West-East distance (km)
Source: Isakov et al. (2007).
36 km CMAQ grid
12 km CMAQ grid
4 km CMAQ grid
Census tract
centroids
in Philadelphia
county:
Tracts in each
4x4km
grid cells are shown
in different colors
Figure 3-72. Grid resolution of the CMAQ model in Philadelphia compared with distribution of
census tracts in which exposure assessment is performed.
For any spatial interpolation models, grid resolution is another source of error addressed. Isakov et
al. (2007) linked CMAQ with the Hazardous Air Pollutant Exposure Model for exposure assessment in
Philadelphia. Their simulation was implemented on a 4 km nested grid within 12 km and 36 Ion grids to
bring the scale of their model from national to urban scale. However, the census tracts in which Isakov et
al. (2007) sought to describe exposure were distributed on a much finer scale (see Figure 3-72). They
were required to supplement the CMAQ model with an Industrial Source Complex Short Term (ISCST)
dispersion model to resolve the subgrid scale behavior. If concentrations were averaged across the cell
instead, Isakov et al. (2007) found that exposures were overestimated by a factor of two. Appel et al.
(2008) noted that their 36 km simulations provided a closer estimate of SO/ aerosol concentration than
did their 12 km nested simulation, which overestimated concentrations. Hogrefe et al. (2007) also noted
overestimation of the CMAQ model at the 12 km scale, where multiple point interpolation was used to
obtain subgrid estimates. Model convergence theory would suggest that the 36 km simulation is not
actually more accurate but coincidentally closer to the physical values (Roache, 1998). Likewise, use of
geospatial statistical methods for grid interpolation, as performed in the SHEDS/MENTOR simulation by
December 2008
3-142
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Georgopoulos et al. (2005), provides another methodology for grid interpolation. Similar to Isakov et al.
(2007), Georgopoulos et al. (2005) were linking CMAQ with an exposure model for estimation of
neighborhood-scale effects. The authors found that CMAQ underestimated PM2 5 concentration at many
times during the simulation.
3.7.3. Findings from PM Exposure Studies
This section is intended to summarize current knowledge regarding exposures to ambient PM in
outdoor microscale environments including streets, sidewalks, and inside vehicles and in indoor
microenvironments such as office buildings and residences where the building's air exchange properties
in conjunction with the physical and chemical properties of the PM dictate levels of exposure to
ambient PM.
3.7.3.1. Outdoor Exposure to Ambient PM
Table 3-28 contains data from recent studies comparing outdoor personal exposure to fixed site
monitors. Only studies where samples were obtained outdoors and compared with a community-based
ambient monitoring site were included because indoor microenvironments have penetration losses that
impact the comparability of the results. Note that some of these studies included enclosed transportation
microenvironments (e.g. cars, buses, subways), but all studies examined personal exposure in the outdoor
microscale environment. Also note that studies must be reviewed cautiously because most used different
instrumentation for personal, microenvironmental, and ambient measurements so that the losses related to
each instrument may vary. The Violante et al. (2006), Kaur et al. (2005a; 2005b), and Adams et al. (2001)
studies showed that outdoor personal exposures to PMi0 and PM2 5 near urban roads was significantly
higher than fixed community-based ambient monitoring site measurements. Curbside measurements
obtained at a fixed site in the Kaur et al. studies were typically lower than exposures during transit
(including during walking and cycling), but in the Adams et al. (2001) study exposures were higher than
curbside during the summer and lower than curbside during the winter. This suggests that particle
retention within the street canyon can lead to elevated local on-street concentrations through trapping of
sources and that this phenomenon may be modified by seasonal effects. Interestingly, Kinney et al. (2000)
showed that on-street PM2 5 concentrations were not significantly different from ambient PM2 5
measurements. However, in this study, EC was shown to increase linearly with increasing traffic counts
with large spatial variations where two sites had concentrations significantly higher than ambient
measurements. These observations suggests caution should be taken regarding the representativeness of
the community averaged monitoring data in Section 3.7.2, as well as hypotheses regarding the impact of
December 2008
3-143
DRAFT-DO NOT QUOTE OR CITE
-------
local sources (e.g. nearby traffic, resuspension from movement of vehicles and people) on personal
exposure.
Table 3-28. Examples of studies comparing outdoor personal exposures with fixed site ambient
concentrations.
Reference
Ambient
monitors
Personal
monitors
Microenvironment,
other variables
Ambient v. Personal
Association
Primary Findings
Violante et
al. (2006)
Bologna,
Italy
Fixed PM10 and
benzene
monitoring
station (method
not specified).
Active pump with
PM10 PEM, passive
sample for benzene
desorbed and
analyzed by GC-MS.
Localized traffic density
(vehicles/h);
Meteorology (wind speed,
wind direction, visibility,
relative humidity).
Personal: 185.10 ± 38.52 |jg/m3
Fixed: 43.56 ± 24.10 |jg/m3
(p<0.0001); small but significant
correlation observed (R2 = 0.19,
p = 0.035) but disappeared after
outlier removal (R2 = 0.09,
p = 0.165).
Fixed PM10 correlated with
multivariate model of traffic
and meteorology but not
personal PM10; relationship
between benzene and PM10
not explored.
Kauret al.
(2005b)
London, UK
Fixed TEOM for
PM2.5 and fixed
CO monitor at
ambient and
curbside sites.
High flow personal
samplers for PM2.5,
P-Trak monitors for
UFP, Langan T15
andT15vfor CO.
Exposures stratified by
mode of transport (walk,
cycle, bus, car, taxi).
Average PM2.5 at TEOMS was 3
times lower than average
personal PM2.5 sample, and 8
times lower than max personal
PM2.5 sample.
PM2.5 exposures during
walking significantly lower than
during car and taxi rides, UFP
exposures during walking
significantly lower than bus
and car rides, cycling
exposures to PM2.5 and UFP
not significantly different from
those on bus, car, or taxi.
Kauret al.
(2005a)
London, UK
Fixed TEOM for
PM2.5 and fixed
CO monitor at
ambient and
curbside sites.
High flow personal
samplers for PM2.5
analyzed
post-sample for
reflectance for EC,
P-Trak monitors for
UFP, Langan T15
andT15vfor CO.
Volunteers walking at set
times and directions along
Marylborne Rd in London.
Fixed vs. personal PM2.5:
slope = 0.29, R = 0.6; personal
PM2.5 measurements were >2
times background levels and
more than 15 |jg/m3 greater
than curbside measurements.
Pedestrian exposures were
significantly higher than fixed
site curbside (or ambient
measurements. Results
indicate that exposure decline
up to 10% from curb-side to
building edge within a street
canyon.
Adams
(2001)
London, UK
Fixed TEOM for
PM2.5 and fixed
CO monitor at
ambient and
curbside sites.
High flow personal
samplers for PM2.5.
Exposures stratified by
mode of transport (cycle,
bus, car, subway).
Median values: (pg/m3)
Summer Winter
Cycle
34.5
23.5
Bus
39.0
38.9
Car
37.7
33.7
Subway
247.2
157.3
Fixed
15
13
Curb
24
37
Exposures were 2.3 - 16.5
times higher than ambient and
1.4-10.3 times higher than
curbside during summer.
During winter, only subway
exposures were appreciably
higher (4.3 times) than
curbside.
Kinney etal. Ambient site
(2000) filter in greased
New York impactorwith
City, NY P™P>
/uQrum\ absorbanc©
testing on filter
for EC.
Three high traffic
sites filter in greased
impactorwith pump;
absorbance testing
on filter for EC.
Traffic counts per hour.
Mean values: (pg/m3)
PM2.5 EC
Site 1 45.7(10.1) 6.2
(1.9)
Site 2 47.1 (16.4) 3.7
(0.6)
Site 3 36.6(10.8) 2.3
(0.9)
Ambient 38.7(10.9) 1.5
(0.5)
PM2.5 at high traffic sites was
not significantly higher than
ambient; EC was significantly
higher than ambient at 2 sites.
EC increased linearly with
traffic counts.
December 2008
3-144
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Outdoor Exposure to Local Sources
Industrial Sources
Industrial sources of aerosols and characterization of their transport and dispersion among exposed
populations have been widely studied. S042 and certain trace metals can largely be attributed to utilities
operations and industrial processes, while N03 and carbonaceous aerosols are derived from a variety of
sources. The 2004 PM AQCD discusses this knowledge extensively, and so the reader is referred there for
detailed information. Distinction of industrial sources from vehicular and indoor sources continues to
pose a challenge in the exposure assessment field (e.g., Nerriere et al., 2007).
Short-Term Sources
Short-term sources, such as time-limited biomass combustion (e.g. forest fire), or sources of rare
origin, such as demolition, can also contribute to acute, high exposures. In a study of air quality following
demolition of a hospital building, Hansen et al. (2008) showed increases of up to four times
pre-demolition concentrations for PM5_i, PMi_0.5, PM0 5.0 3, and UFP concentrations immediately following
demolition. Likewise, Ng et al. (2005) published a time series of PM2 5 concentration in New York City
from September to November, 2001, which includes the World Trade Center collapse. Significant spikes
in concentration were observed during that incident, as shown in Figure 3-73. The Ebelt et al. (2005)
study also illustrates the issue of non-representative PM events. On one day dust from the Gobi Desert
caused an increase in the concentration of fine and coarse PM. When this day was deleted from the
analysis, the relationships were changed and the associations, especially for coarse PM, became larger
and more significant. Forest fires, described in more detail for PRB in section 3.6, have shown spikes in
measured concentrations (Generoso et al., 2007; Mathur, 2008; Wang et al., 2006a) and signatures in
source apportionment exposure studies (Ogulei et al., 2006).
December 2008
3-145
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
195
—— MVPO (all year) —Hospital (since 09/14)
— Hunter (all year) -"-PS64 (all year)
—RaceU (since 09/25)
Hunter College
2 130
MVPO
Pa cell
NYU Hospital
Ot— t-0CNJ-t-0CM-*-t-0*-'*j-r<-o>
OO-«-*-'--»-CN|CMtNC0OOO'r-"«-*-'<-CvJCNI(NC0OOO-*-*-CMCNCNC>J
0>CT10)C3>CnO>CT>a>0>0500000000000-<-T-T-'^'r-T-T-T--^
Month- Day-Time
Source: Ng et al. (2005).
Figure 3-73. Time series plot of PM2.5 concentration measured at various sites during September -
November 2001. The Hunter College and PS 64 monitors operated near Lower
Manhattan during the World Trade Center collapse, and the MVPO monitor operated in
Upper Manhattan. The Pace and Hospital monitors began sampling several days to
weeks after the collapse,
The Near-Road Microscale Environment
Sections 3.3 and 3.5 describe the physical and chemical composition of traffic emissions as well as
evolution of the plume away from traffic sources. Particle chemistry is an important determinant of
human exposure because the chemical constituents dictate the toxicity and size-selective behavior of the
PM exposure, which then affect the mode and location of deposition of particles in the respiratory system.
The increased magnitude of the sources, turbulent transport, and physical and chemical transformation
processes described in Sections 3.3 and 3.5 in the near-road region complicate exposure assessment for
those living in the near-road environment (Zhang et al., 2005b; Zhou and Levy, 2007; Zhu et al., 2002b).
Farmer et al. (2003) found that exposure to particle-bound B[a]P and PAH can be 2-3 times higher among
those routinely exposed to outdoor traffic emissions (e.g. police, bus drivers) compared with control
subjects.
In-Vehicle and In-Transit Exposures
In-vehicle pollution has been identified in various studies as a source of exposure to PMi0, PM2 5,
and ultrafine PM (Briggs et al., 2008; Diapouli et al., 2007; Fruin et al., 2008; Gomez-Perales et al., 2004;
December 2008
3-146
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
2007; Gulliver and Briggs, 2004, 2007; Rossner et al., 2008; Sabin et al., 2005). Results from recent
studies are provided in Annex A. In many of these studies, in-vehicle exposures are shown to be
comparable or less than that of walkers on the same route. Typically, in-vehicle exposures were still
higher than community-based ambient monitor concentrations for TSP and PMi0. However, as particle
size decreased to the fine and ultrafine range, less distinction between in-vehicle and ambient
concentrations were observed for PM mass or count, with the exception of the Diapouli et al. (Diapouli et
al., 2007) study where in-bus concentrations were several times higher than indoor or outdoor residential
and school concentrations. Fruin et al. (2008) and Westerdahl et al. (2005) observed that in-vehicle
concentrations increased for freeways in comparison with arterial roads. However, Sabin et al.
demonstrated for school buses that emissions control technologies had a significant impact on in-bus
concentrations. Although not tested here for other vehicle types with respect to PM, this finding suggests
that some in-vehicle emissions are due to self-pollution. Behrentz et al. (2004) tested self-pollution with
school buses using SF6 tracer gas and demonstrated that as much as 0.3% of in-vehicle air comes from
self-pollution, and that this number was roughly ten times greater than in-vehicle concentrations related to
self-pollution on newer buses. The Behrenetz et al. study also measured EC and particle-bound PAH and
found that 25% of the variability in EC concentration was related to self-pollution, defined by Behrentz et
al. (2004) as the fraction of a vehicle's own exhaust entering the vehicle microenvironment. These
findings are important for partitioning local and ambient sources of pollution during transport in vehicles
for exposure estimation. Based on Fruin et al.'s (2008) estimation of 1.5 hours spent in vehicles each day
(for the Los Angeles area, but this value is typical for many other large metropolitan areas), cumulative
in-vehicle exposure can become important. Mixed findings from these studies suggests that in certain
limited cases, e.g., arterial road travel with low traffic, in-vehicle concentrations of fine particles can be
reasonably represented by community ambient monitors.
3.7.3.2. Indoor and Average Personal Exposure to Ambient and Non-Ambient PM
A number of exposure studies have been published since 2002. Annex A lists those studies
performed in the U.S. by region of the country with personal, microenvironmental, and ambient mass
concentrations presented (note that chemical speciation data where available are presented and discussed
below). The majority of these studies present PM2 5 concentration data for personal exposure and ambient
concentrations. Some of these studies present ambient concentration as that measured outside the test
building, while others use a community site monitor. As would be expected, there is considerable
variability within and across regions of the country with respect to indoor exposures and ambient
concentrations. Furthermore, some regions are represented by only one or two studies, while some
regions have many studies. Among these regions, most are represented by only one or two metropolitan
December 2008
3-147
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
areas. For this reason, the results presented may not be broadly representative. These differences highlight
the uncertainties surrounding various estimates of the ambient contribution to personal exposure. This
variation can be attributed to a number of factors, including scope and magnitude of microenvironmental
sources, proximity to microenvironmental sources, ambient concentrations of PM, percentages of time
spent in various microenvironments, the age and condition of indoor microenvironments, and outdoor
meteorology. Findings related to source apportionment, infiltration, and differential infiltration studies are
discussed further in the following subsections.
Estimating the Ambient Component of Persona/ Exposure
In the context of determining the effects of ambient pollutants on human health, the association
between the ambient component of personal exposures and ambient concentrat88ions is more relevant
than the association between personal total exposures (ambient component + nonambient component) and
ambient concentrations. If there are no indoor or other non-ambient sources of a pollutant, the total
personal exposure is equal to the ambient personal exposure. However, indoor or other non-ambient
sources could significantly affect personal exposures to many pollutants. Strand et al. (2007) reviewed
methods for estimating the ambient component of personal exposure and their limitations.
Wilson et al. (2000) first proposed that sulfate could be used as a tracer of the ambient PM
infiltration rate. Sarnat et al. (2002) also noted that it is reasonable to assume that the size distribution of
ambient S04 particles is sufficiently similar to the size distribution of ambient PM2 5, and therefore that
the ambient S04 to personal S04 ratio is an acceptable surrogate for the ratio of the ambient PM25
exposure to the ambient PM2 5 concentration. Sulfate has been used this way by, e.g., Ebelt et al. (2005),
Wallace and Williams (2005), and Wilson and Brauer (2006). For this method to be successful, indoor or
other non-ambient sources of the tracer must be small compared to ambient sources over the period of
sampling. Wilson and Brauer (2006) noted that environmental tobacco smoke and tap water used in
showers or humidifiers are indoor sources of sulfate. Other concerns in using sulfate as a tracer for PM2 5
arise because sulfate tends to be concentrated in smaller particles and thus it might be a better tracer for
fine mode particles than for all PM2 5 particles. Strand et al. (2006) suggested that Fe be used as an
additional tracer to correct for the infiltration of larger PM2 5 particles. In their study, they noted that
indoor sources of Fe were small. However, in other environments there could be more substantial
contributions from tracking iron in soil indoors. The spatial variability of Fe is also larger than that of
PM2 5 across urban areas. Volatilization of nitrate or organic compounds after infiltration of PM2 5 indoors
could lead to bias in exposure estimates (Lunden et al., 2003a). This could be a large problem in areas in
which PM contains a large semi-volatile component.
December 2008
3-148
DRAFT-DO NOT QUOTE OR CITE
-------
1 Figure 3-71 shows total exposure to S04 as a function of measured ambient S04 concentration.
2 Figure 3-75 shows estimated ambient exposure to PM2 5 as a function of measured ambient PM2 5
3 concentration, where ambient personal exposure is calculated from the ambient exposure factor for S04.
4 Close agreement between these figures can be observed. Figure 3-76 shows total exposure to PM2 5 as a
5 function of measured ambient PM2 5 concentration. However, the total exposure to PM2 5 shows virtually
6 no association with ambient PM2 5 because it contains non-ambient contributions to PM2 5.
T$o*= 0.74 C«>4-0.01
R2 = 0.75
0
1
2
3
5
6
Ambient Concentration, (pg/m3)
Source: Wilson and Brauer (2006).
Figure 3-74. Total exposure to SO4 as a function of measured ambient SO4 concentration.
0.76 Cli- 0.093
R*~= 0.62
20
Ambient Concentration, C2 S ((jg/rrr3)
Source: Wilson and Brauer (2006).
Figure 3-75. Estimated ambient exposure to PM2.5 as a function of measured ambient PM2.5
concentration.
December 2008
3-149
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
£ 100
3.
« ao
&
3 60
in
W 40
CO
| 20
CL
"ra
¦'
72.5 =
R2.S
= 0.77 C2 5 + 8.24
= 0.07
¦ 2
~
~
~
I
~ *.
~ ~ *
X
~
~ ~ ~
*
~ ^
~
~
0 5 10 16 20 25 30
Ambient Concentration, C2 5 (pg/m3)
Source: Wilson and Brauer (2006).
Figure 3-76. Total exposure to PM2.5 as a function of measured ambient PM2.5 concentration.
The estimated ambient exposure to PM2 5 is well correlated with measured ambient PM2 5
concentration with zero intercept implying that non-ambient sources were unimportant. This technique
works well in areas where sulfate is a regional pollutant insuring that its spatial variations are small and it
is highly correlated with PM2 5 (Kim et al., 2005a; U.S. EPA, 2004). The inferences drawn from this
method may still apply in areas where sulfate is a minor component of PM2 5 or where there are
significant non-ambient sources of sulfate as long as the factors affecting a person's exposure are similar.
Source apportionment techniques could also be used, in principle, to derive ambient personal PM2 5
concentrations. They would be especially useful in areas where the application of a tracer method might
be problematic. Hopke et al. (2003) used PMF to derive source contributions that affected community,
outdoor, indoor samples at a retirement facility in Towson, MD. Strand et al. (2007) noted that the four
outdoor factors (nitrate-sulfate, sulfate, OC, motor vehicle exhaust) would constitute an estimate of the
personal ambient PM2 5 concentration. However, the data used in this portion of the analysis were
obtained only with fixed monitors and did not include measurements made by personal exposure monitors
(PEMs). They also used the Multilinear Engine (ME) to derive factors that were required to contribute
jointly to central indoor and outdoor, individual apartment and individual PEM samples of a panel of
residents. Hopke et al. (2003) found three sources (sulfate, unknown [perhaps combustion related] and
soil) jointly contributing 46%, 13% and 4% of PM2 5 to the PEM samples. Further source resolution was
not possible because of a lack of data for a number of components in the PEM samples. The largest and
most clearly identified contribution to personal exposure was from the sulfate factor. This study also
determined that a few minor indoor and personal activity sources contributed < 10% of the ambient
sulfate source to personal exposures. Although this study did not directly attempt to determine the
ambient component of personal exposures, the sulfate factor could represent a lower limit on the ambient
December 2008
3-150
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
personal concentration, as the study was conducted in an area where ambient sulfate concentrations are
fairly homogenous and where they are highly correlated to PM2 5. This obviates the limitation from not
using data from the ambient monitors reporting to AQS.
The correlations between personal ambient PM exposures and ambient PM concentrations in
different types of exposure studies are relevant to different types of epidemiologic studies. There are three
types of correlations generated from the different study designs: longitudinal, "pooled," and daily-average
correlations as described in the 2004 PM AQCD and later in the 2008 NOx ISA (U.S. EPA, 2008c) and
the 2008 SOx ISA (U.S. EPA, 2008d). Generally, strong associations between personal exposures and
ambient concentrations were reported in the longitudinal studies. Wilson and Brauer (2006) and Ebelt et
al. (2005) reported that the Pearson correlation coefficients between personal ambient exposure
(estimated by the sulfate tracer method) and ambient concentration of PM2 5 (for individual subjects)
ranged from 0.77 to 0.92, with a median of 0.88 in the Vancouver exposure study (16 COPD subjects, and
seven repeated measurements for each subject). Koutrakis et al. (2005) reported the median Spearman
correlation coefficients between personal sulfate exposure and ambient sulfate concentration were above
0.60 during both winter and summer in Boston and Baltimore (15 subjects with 12 consecutive
measurements during each season in both Boston and Baltimore). For another Baltimore cohort (15 senior
subjects with up to 23 consecutive measurements for each person), Hopke et al. (2003) reported that the
median Pearson correlation coefficient between personal exposure to sulfate factor and ambient sulfate
factor was 0.91 (ranging from 0.56 to 0.95 for different subjects), while the median Pearson correlation
coefficients were 0.34 for the crustal factor (ranging from -0.05 to 0.62), and 0.31 for an unnamed factor
(ranging from -0.01 to 0.88), respectively.
Wilson and Brauer (2006) reported that the pooled Pearson correlation coefficient was 0.79 for
personal ambient exposures (estimated by tracer element method) vs. ambient concentrations, and was
0.001 for personal nonambient exposure vs. ambient concentrations.
Strand et al. (2007) conducted an exposure study in Denver (from 2002 to 2004) for 6-12 year old
school children. Up to 10 personal exposure samples were collected on each day, and ambient
concentrations were measured simultaneously at a fixed site located at the school. The daily average
personal sulfate exposure was strongly associated with ambient sulfate concentration (r = 0.96, 120 > N >
100).
Exposure to PM Species and Source Apportionment of Ambient and
Non-Ambient Contributions to PM
Annex A presents exposure studies that include chemical speciation data. Some of these studies
focused on S042 , N03 or carbonaceous aerosols (EC, OC, particle-bound PAHs, B[a]P), while others
December 2008
3-151
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
measured concentrations of trace metals from crustal (Ca, Fe, Mn, K, Al, S, CI in salt), traffic (Al, Ca, Fe,
K, Mg, Na, Ba, Cr, Cu, Mn, Ni, Pb, S, Ti, V, and Zn; or industrial (particle-bound Hg, CI, V, Zn, Ti, Cu,
Pb) sources. A number of source apportionment studies have been performed over the last five years to
determine the contribution of outdoor sources to indoor and personal PM constituents. These are listed in
Annex A.
A number of studies have examined exposure to S042 (e.g., Brunekreef et al., 2005; Ebelt et al.,
2005; Kim et al., 2005b; Koutrakis et al., 2005; Noullett et al., 2006). Hopke et al. (2003) and Zhao et al.
(2006b) showed that secondary S042 provides the largest ambient contribution to personal and indoor
exposures. These studies took place on the east coast in Baltimore and Raleigh/Chapel Hill, NC. In
Larson et al.'s (2004) source apportionment study in Seattle, vegetative burning was the most significant
source of outdoor origin. Zhao et al. (2007) performed a source apportionment study of personal exposure
to PM2 5 among residents in Denver and also saw lower contributions from secondary S042 in
comparison with motor vehicle emissions and secondary N03~. This suggests that personal exposure to
S042 in parts of the west is lower than in the mid-Atlantic.
Source apportionment for carbonaceous aerosols is complicated by the fact that they can be derived
from indoor and outdoor combustion sources. Carbonaceous aerosols are difficult to trace to specific
indoor and outdoor sources because combustion is widespread. Sorcnson et al. (2005), Ho et al. (2004),
Larson et al. (2004), and Jansen et al. (2005) all found that personal and microenvironmental exposure to
total or BC was lower than that measured outdoors, while Sarnat et al. (2006a) showed significant
associations between personal and ambient measurements of EC for measurements taken during the fall
(slope = 0.66-0.73) and summer under high ventilation conditions (slope = 0.41). Wu et al. (Wu et al.,
2006), Delfino et al. (2006), and Turpin et al. (2007) all demonstrated much higher levels of OC
compared with EC in personal samples, possibly due to indoor sources of OC from cooking and home
heating. Reff et al. (2007) and Meng et al. (2007a) both reported findings from the RIOPA study in Los
Angeles, Houston, and Elizabeth, NJ. Results from Reff et al. (2007) are shown in Figure 3-77. These
reveal significantly higher detection of aliphatic C-H functional groups indoors and in personal samples
compared with outdoors. This information can be used to distinguish indoor and outdoor carbon in future
source apportionment studies. Little regional variation in the aliphatic, carbonyl, or S042 groups tested
were reported in this study. In Meng et al., (2007a) indoor exposures were shown to decrease for
secondary formation aerosols including S042 but excluding N03 (not tested) when compared with
outdoor concentrations, and indoor exposures to mechanically generated aerosols increased in comparison
with outdoors. Differences in infiltration based on constituents are shown in Figure 3-78.
December 2008
3-152
DRAFT-DO NOT QUOTE OR CITE
-------
Los Angeles Co., CA
^ c?• C!
^ ^ ^ rk ^
a
o o* o5*
^ cP
Elizabeth, NJ
E 20
P 10
I
u
Ov
^ ^ ^ Cf
°
_6 -°
<55
Houston, TX
a
^ ^ d> «/ ,
_0S rP* rf rS
J? •Pjf.jf.l?
¦>©
o
\<>
rSy
a
o* o** o**
,<=r <3U
-><£
Source: Reffet al. (2007).
Figure 3-77. Apportionment of aliphatic carbon, carbonyl, and SO42 components of outdoor,
indoor, and personal PM2.5 samples, for Los Angeles (top), Elizabeth (center), and
Houston (bottom).
December 2008
3-153
DRAFT-DO NOT QUOTE OR CITE
-------
o
o
'O
c
30 ¦
20 -I
10 Y
Mechanical Fin( = 0.04
y = 0.04x +0.51, Ra = 0.51
/
/
/
/
/
/
/
/
/
/ • AER: < 0.5 h"'
V AER: 0.5 -1.5 h"1
/ ~ AER: > 1,5 h'1
/
- v
~
|3D-
10 15 20 25 30 35
30
50
C
O
25 ¦
20 -
15 •
Primary Fln( = 0.51
y = 0.51x + 0.19, R2 = 0.61
/
/
cd 10 -
C
1.5 h"1
30
40
50
Outdoor Concentration (jig/m1)
Source: Meng et al. (2007a).
Figure 3-78. Apportionment of infiltrated mechanically-generated (top), primary combustion
(center), and secondary combustion (bottom).
December 2008
3-154
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Trace metal studies have shown variable results regarding personal exposure to ambient
constituents. For instance, Molnar et al. (2006) found that personal exposure was higher than outdoor and
ambient concentrations for mostly crustal CI, K, Ca, Ti, Fe, and Cu. However, Adgate et al. (2007) found
that personal exposures were higher than ambient for Fe, Mg, K, Zn, Cu, Pb, and Mn but lower than
ambient for Al, Na, and Ti. Larson et al. (2004) found that personal exposure to Ca and CI were higher
than concentrations measured at ambient (central site) and residential outdoor monitors, lower for Fe, K,
Mn, and As and the same for Al, Br, Cr, Cu. Source apportionment for trace metals can vary significantly
among cities and over seasons. For instance, in a Baltimore source apportionment study, exposure to Mn
could be attributed nearly equally to the Quebec wildfires, roadway wear, and soil, while Pb exposure was
largely found to be due to a local incinerator (Ogulei et al., 2006). In this case, the Quebec wildfires were
a transient episodic source while roadway wear and incineration were continuous. However, in Larson et
al. (2004), Mn and Pb exposures in Seattle were largely attributable to mobile source emissions, also
stationary sources. For this reason, source composition behavior cannot be assumed for characterizing
exposures and resulting health effects in specific locations or times.
Infiltration and Differential Infiltration
Fjnf varies substantially given a vast array of conditions, and it can best be modeled dynamically
based on a distribution of air exchange and deposition or other ultrafine, accumulation mode, fine, and
coarse PM loss rates rather than a single value (Bennett and Koutrakis, 2006). Given that air exchange
rates within a building vary as a function of temperature and pressure, Fmf is subject to seasonal and
regional changes (Meng et al., 2005a; Sarnat et al., 2006b). These factors make Finf a more accurate
descriptor of infiltration than a simple I/O ratio. This complex term becomes even more complicated
when one considers transformation of the size distribution and chemical composition of the PM through
site reactions, agglomeration, growth, and evaporation given that Fmf depends on particle size (Keller and
Siegmann, 2001). Fmf for PM is influenced by physical mechanisms, such as Brownian diffusion,
thermophoresis, and impaction, all of which are functions of particle size (Bennett and Koutrakis, 2006;
Tung et al., 1999). These differential effects are summarized below. Recent studies on infiltration are
summarized in Annex A. Fmf and I/O are listed where available, although it is recognized that I/O is not as
meaningful a descriptor but provides a sense of the data where Finf has not been calculated.
A number of studies have examined the impact of season on PM infiltration. Season is important
because it impacts the ventilation practices used (e.g. open windows, air conditioning or heating use) and
the ambient temperature and humidity conditions affect the transport, dispersion, and size distribution of
the PM. Pandian et al. (1998) found that residential air exchange rates vary by season as: summer
> spring > winter > fall with summer air exchange roughly 1.5-2 times greater than average air exchange
December 2008
3-155
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
rate for the entire year because they are driven by home air conditioning and heating usage. Allen et al.
(2003) gave information on the range and distribution of Fmf at 44 residences in Seattle. The mean Fmf
(± SD) measured by light scattering for all sampling days was 0.65 ±0.21. Differences in infiltration were
observed for the heating season (0.53 ± 0.16), when windows would be expected to be closed, versus
non-heating season (0.79 ± 0.18). Residences with open windows had a mean Fmf of 0.69 vs. 0.58 for
residences with closed windows. The authors combined the light scattering results with indoor and
outdoor sulfur measurements to estimate that 79 ± 17% of indoor PM2 5 was generated outdoors. This
study provides important data on the distribution of residential Finf values and illustrates the magnitude of
the effect of season and window position on infiltration rates. Barn et al. (2008) and Baxter et al. (2007a)
both also noted that window opening was an important variable. Barn et al. (2008) found Finf of
0.61 ± 0.27 for 13 homes during summer and 0.27 ±0.18 for 19 homes during winter in Canada.
Likewise, location could impact residential ventilation practices and infiltration. Pandian et al. (1998)
observed that residential air exchange rates vary by region as: southwest > southeast > northeast
> northwest, which reflects regional use of air conditioning. Sarnat et al. (2006b) noted differences in
infiltration between coastal and inland residences, although variability in these datasets made the
differences not statistically significant.
Differential infiltration as a function of particle size has been observed to occur. Infiltration factors
for several particle diameters ranging from 20 nm to 10 ^m were reported in Boston by Long et al. (2001)
during summer and fall for nighttime periods, when personal activity patterns would be less likely to
generate indoor PM. The maximum infiltration factor was reported for particles between 80 and 500 nm,
to range from 0.8-1.0. Summer values were uniformly higher than fall values, consistent with higher
observed air exchange rates. The infiltration factor decreased with size above 500 nm, reaching 0.1-0.2
for 6-10 (mi diameter particles. Particles smaller than 80 nm also were reported to have lower infiltration
factors. This demonstrates the size dependence of PM infiltration, which has been further studied by
recent investigators. Sarnat et al. (2006b) examined infiltration as a function of particle size and found
that F^ does vary by particle diameter, as measured by a SPMS-APS system to estimate particle volume.
Figure 3-79 presents Fmf values for size fractions ranging from 0.02-10 (mi. The maximum infiltration
factors were observed around the accumulation mode (0.1-0.5 |im). with Fmf = 0.7-0.8. Reduced
infiltration was observed for coarse-mode particles (0.1-0.2 for Dp = 5-10 (jm) and, to a lesser extent,
ultrafine particles (0.5-0.7 for Dp = 0.02-0.1 (mi). This is consistent with increased removal mechanisms
for those size fractions: settling for coarse-mode particles and diffusion for ultrafine particles.
December 2008
3-156
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
n 0.6
o 0.5
0.01
0.1 1
Particle Size(^m)
10
Source: Sarnat et al. (2006b).
Figure 3-79. Fmf as a function of particle size.
A number of chemical factors influence the tendency for differential infiltration in PM. Lunden et
al. (2003b) studied infiltration of BC and OC aerosols and found that Fmf can also vary substantially as a
function of gas transport properties within differing air exchange rates. This study and Sarnat et al.
(2006b) also showed that BC aerosol infiltration is considerably higher than infiltration of OC, and that
carbonaceous aerosol infiltration differed substantially from NO;, and S042 aerosols under the same
building air exchange conditions. The evidence indicates that the ambient composition differs from the
indoor composition due to differences in penetration. As shown in Figure 3-80, the composition of indoor
PM that has infiltrated from outdoors is different from that of outdoor PM. In this case, the particles
containing photochemical products (probably mostly in the accumulation mode) have a higher infiltration
rate than the larger (probably mostly coarse mode) mechanically generated particles or the smaller
primary combustion particles (probably mostly ultrafine particles in the nucleation or Aitkin nuclei
mode). Hence there is a greater reduction in the concentration of the larger and smaller particles resulting
from losses during penetration and due to deposition within the home.
December 2008
3-157
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
Primary
Photochemical Combustion
Secondary Particles
(~ Accumulation Mode) (~ Ultrafine)
Mechically
Generated
(~ Coarse Mode)
Source: Meng et al. (2007a).
Figure 3-80. Results of the positive matrix factorization model showing differences in the
composition of outdoor PM and PM that has infiltrated indoors.
PM species enriched in the smaller end of the size distribution, such as S042 , will infiltrate more
efficiently than components with larger size distributions, such as iron (Strand et al., 2007). Lunden et al.
(2008) also compared I/O ratios for PM2 5, total carbon, OC, and BC in an unoccupied house and found
the lowest ratio for PM2 5 (0.41 ± 0.2), the highest for BC (0.61 ± 0.2), and intermediate values for total
carbon (0.50 ± 0.2) and OC (0.47 ± 0.2). The authors attributed the lower PM2 5 ratio to indoor loss of
NH4NO3 aerosol. The authors note that their BC ratio of 0.6 is somewhat lower than BC ratios measured
in occupied spaces (Polidori et al., 2006). Conversely, indoor sources in occupied residences contribute to
observed OC I/O ratios greater than 1 in other studies (Polidori et al., 2006; Sawant et al., 2008).
Analytical results for PM2 5 components from the Baxter (2007b) study found Finf of 0.95 ± 0.07 for S and
0.60 ± 0.04 for V, the two components identified as having no indoor sources and which had I/O ratios
significantly less than 1 (Baxter et al., 2007b). The difference between these two values was not
addressed, although association of V with larger particles of lower penetration efficiency could contribute
December 2008
3-158
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
to a lower infiltration rate. Meng et al. (2005a) also noted that the lack of indoor sources of S and V result
in much lower variability in penetration and loss rates.
The resulting composition differences between indoor-ambient PM and outdoor-ambient PM may
also result in differences in toxicity. N03~, a more important PM component in the western U.S., has a
decreased Fmf due to volatilization of N03 indoors. Sarnat et al. (2006b) calculated Finf values for N03 .
PM2 5, and BC, and found the values to increase in that order. N03 was low (median = 0.18,
IQR = 0.12-0.33), while BC was high (median = 0.84, IQR = 0.70-0.96); the intermediate value of PM2 5
(median = 0.48, IQR = 0.39-0.57) reflected its composition as a mixture of those two components (among
others). Indoor volatilization of N03 enriches ambient PM in other components, creating differences in
toxicity between indoor ambient and outdoor ambient PM. The high infiltration of non-volatile BC
creates additional sorption sites for organics, including indoor-generated compounds. Meng et al. (2007a)
found that secondary formation accounts for 55% of indoor aerosols of outdoor origin, while primary
combustion accounts for 43% and mechanical generation for 2%. Meng et al. (2007a) noted that
secondary formation processes often result in slightly larger particles so that diffusion losses are not as
great as for primary combustion particles, composed primarily of nucleation and condensation modes.
Likewise, Polidori et al. (2007) suggest that similarities in the EC and OC size distributions and
infiltration factors indicate low vapor pressure secondary organic aerosols in the OC component.
Variations in the presence of outdoor PM indoors relates to the species composition and makeup of PM.
Differences in infiltration and in removal behavior once indoors tend to cause differences in the species
composition. The toxicological profile of outdoor and indoor-penetrated outdoor PM would therefore be
expected to be different. An example of these studies is Molnar et al. (2007), who reported indoor/outdoor
ratios of 0.4-0.7 for residences, schools, and preschools in Sweden, using sulfur and lead as tracers for
ambient PM. These ratios are consistent with those observed previously.
Exposure to PM in Copollutant Mixtures
Analysis of personal exposure to multi-pollutant mixtures is an area of growing research. Several
multi-pollutant studies involving ultrafine, fine, and coarse PM are presented in Table A-91.
Understanding the health impacts of complex multi-pollutant mixtures, including multiple PM species,
can have a substantial impact on the interpretation of health effects data. Challenges are presented in
accurately estimating the components of a mixture, their concentrations, and personal exposure to those
species.
One question that has been raised is whether copollutants act as confounders in PM exposure
assessments. Sarnat et al. (2001) explored the relationship between PM and copollutant gases and
suggested that certain gases can serve as surrogates for describing exposure to other air pollutants. Sarnat
December 2008
3-159
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
et al. (2001) found significant associations between personal exposure to PM25 and ambient
concentrations of 03, N02, CO (significant only for winter), and S02 in a panel study conducted in
Baltimore. Personal exposures to PM2 5 and personal exposures to the gases were not correlated in this
study. This result may have arisen in part because personal exposures to the gases were often beneath
detection limits of the personal monitoring devices. Significant associations were found for ambient PM2 5
and ambient 03, N02, wintertime CO, and S02 concentrations. In both cases, ambient 03 was positively
associated with PM2 5 in summer and negatively associated with PM2 5 in winter, which is consistent with
findings of Ito et al. (2005) for ambient PM (using PM2 5 for New York and Philadelphia and PMi0 for
Houston, Minneapolis-St. Paul, Detroit, Cook County, IL (Chicago), St. Louis) and ambient 03 for all
cities but Houston. Sarnat et al. (2001) suggested that relatively high correlations between ambient PM2 5
and ambient N02 may be due to the emission of both substances by motor vehicles. Schwartz et al. (2007)
also used data from the Baltimore panel study to simulate distributions of personal exposures and ambient
concentrations of PM2 5, PMi0, S042 . N02, and 03. They found that personal PM2 5 was significantly
associated with ambient N02 and ambient 03 (in an inverse relationship) and personal S042 was
significantly positively associated with ambient 03. Brook et al. (2007) also noted the correlation between
ambient PM2 5 and ambient N02 in a study of ten Canadian cities but suggested that N02 is a better
indicator of PM2 5 exposure than ambient PM2 5 because it is less variable. In a subset of this work, Brook
et al. (2007) also showed that the correlation between ambient N02 and ambient hopanes, indicative of
vehicle exhaust, was stronger than the correlation between ambient PM2 5 and ambient hopanes (these
data were obtained for two weeks in Windsor, ON in March 2001).
Seasonality of the associations could be a result of seasonal variability in photochemistry, source
generation, and building ventilation. Sarnat et al. (2005) observed associations between personal PM2 5
and ambient 03, N02, and S02 exposure for a group of healthy senior citizens and school children in
Boston for summer but not for winter. In this study, correlations between personal exposure to ambient
PM2 5 and personal 03 exposures were observed, unlike the Baltimore study. In their study of personal
exposure to ambient air pollutants in Steubenville, OH, Sarnat et al. (2006a) found that, in the summer,
low but significant associations existed of ambient 03 with personal PM2 5, S042 , and EC. In the summer,
low but significant associations between ambient S02 and personal PM2 5 and between ambient N02 and
personal EC were also observed. In the fall, ambient 03 had a weak but significant association with
personal EC, and S02 had a weak but significant association with personal S042 . Ambient N02 was
significantly associated with personal PM2 5, S042 . and EC with somewhat higher coefficient of
determination (R2 = 0.25-0.49). In summary, the evidence is mixed that ambient gases can be considered
surrogates of PM2 5 exposure. There is evidence that associations between ambient gases and personal
exposure to PM2 5 of ambient origin exist but are complex and vary by season and region.
December 2008
3-160
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
3.7.4. Exposure Assessment and Socioeconomic Status
The number and complexity of environmental equity studies have increased dramatically since the
First National People of Color Environmental Leadership Summit was held in 1991. The environmental
equity movement was borne from this conference and Executive Order 14898, in which President Clinton
declared that equity issues must be factored into the Environmental Impact Statement process. The
preponderance of environmental equity studies related to air quality has not examined the relationship
directly between airborne PM and SES because the scale of FRM networks for monitoring PMi0 and
PM2 5 (or other criteria air pollutants) in many metropolitan areas is much larger than the spatial variations
in SES across neighborhoods. This point is illustrated in Figure 3-81, which shows the 15 km radius
around each PM2.5 monitor within the Philadelphia CSA. (Note that maps of numbers below poverty level
per square mile and numbers with less than a high school diploma per square mile compared with PMi0
and PM2 5 sampler density are provided for each of the fifteen CSAs/CBSAs in Annex A). Here, it can be
seen that for areas within 15 km of a PM2 5 monitor in the Philadelphia CSA, significant variability in SES
exists. Tables 3-29 through 3-32 show percentages within the 1 km, 5 km, 10 km, and 15 km of the
monitor for populations below the poverty level and with less than high school, high school or more, and
college graduate education levels for the fifteen CSAs/CBSAs. These tables illustrate that the monitoring
stations do provide some coverage related to low SES groups. For PM10, 27-94% of the population below
poverty level and 41-92% of the population with less than a high school education are within 15 km of a
monitor. 47-93% of the population below poverty level and 44-91% of the population with less than high
school education for those are within 15 km of a PM2 5 monitor for the fifteen CSAs/CBSAs studied.
However, neighborhood scale SES issues are not shown to be we 11-represented, with 0.2-6% of the
population below poverty and 0.2-5% of the population with less than a high school education within 1
km of a PMio monitor. Likewise, 0.4-7% of the population below the poverty level and 0.4-5% of the
population with less than high school education are within 1 km of a PM2 5 monitor.
Annex A lists 34 recent studies that specifically include the association between ultrafine,
accumulation, PM2 5, or PMi0 and a health outcome as modified by one or more SES variables. The
effects of SES as an effect modifier in health effects studies are discussed in Section 8.2.4.
December 2008
3-161
DRAFT-DO NOT QUOTE OR CITE
-------
Table 3-29. Proximity to PM10 and PM2.5 monitors among adults older than 25 with less than a high
school education by city. Percentages are given with respect to the total population
per city provided.
Region
Total CSA/CBSA
<1 km
<5 km
<10 km
< 15 km
N
N
%
N
%
N
%
N
%
PROXIMITY TO PM10 MONITORS
Atlanta
397037
5828
1.47
69214
17.43
157705
39.72
222027
55.92
Birmingham
139841
7135
5.10
60727
43.43
83681
59.84
93142
66.61
Boston
360791
5895
1.63
86575
24.00
156912
43.49
189101
52.41
Chicago
934491
8405
0.90
100953
10.80
278411
29.79
531905
56.92
Denver
170110
5094
2.99
53462
31.43
107939
63.45
143914
84.60
Detroit
464605
3428
0.74
53780
11.58
148470
31.96
253490
54.56
Houston
635036
6529
1.03
141564
22.29
367943
57.94
468206
73.73
Los Angeles
1963819
7114
0.36
265004
13.49
869291
44.27
1496953
76.23
New York
n/a
rJa
rJa
n la
n la
n/a
n/a
n la
rJa
Philadelphia
598949
5045
0.84
46708
7.80
81712
13.64
163386
27.28
Phoenix
383484
23819
6.21
199483
52.02
316163
82.44
359337
93.70
Pittsburgh
256990
11379
4.43
100417
39.07
148283
57.70
174178
67.78
Riverside
477430
11257
2.36
151939
31.82
328305
68.77
381372
79.88
Seattle
254781
492
0.19
26330
10.33
73186
28.72
109413
42.94
St. Louis
264721
8756
3.31
86139
32.54
162996
61.57
186303
70.38
PROXIMITY TO PM2.S MONITORS
Atlanta
397037
1816
0.46
44972
11.33
193590
48.76
280777
70.72
Birmingham
139841
2683
1.92
37608
26.89
95116
68.02
108593
77.65
Boston
360791
24546
6.80
184245
51.07
247784
68.68
294008
81.49
Chicago
934491
24519
2.62
400070
42.81
769226
82.32
869320
93.03
Denver
170110
4018
2.36
53903
31.69
120088
70.59
151516
89.07
Detroit
464605
10931
2.35
178094
38.33
352951
75.97
396954
85.44
Houston
635036
2581
0.41
47042
7.41
195735
30.82
301048
47.41
Los Angeles
1963819
23815
1.21
489939
24.95
1346707
68.58
1768102
90.03
New York
2450252
171684
7.01
1385259
56.54
1978746
80.76
2232024
91.09
Philadelphia
598949
26912
4.49
366191
61.14
493003
82.31
533423
89.06
Phoenix
383484
8294
2.16
100266
26.15
197412
51.48
246594
64.30
Pittsburgh
256990
8008
3.12
90904
35.37
161286
62.76
197209
76.74
Riverside
477430
9768
2.05
131733
27.59
276475
57.91
320506
67.13
Seattle
254781
1510
0.59
30561
12.00
87832
34.47
128641
50.49
St. Louis
264721
8550
3.23
87827
33.18
183985
69.50
210177
79.40
December 2008
3-162
DRAFT-DO NOT QUOTE OR CITE
-------
Table 3-30. Proximity to PM10 and PM2.5 monitors for the total population under poverty line by
city. Percentages are given with respect to the total population per city provided.
Region
Total CSA/CBSA
<1 km
<5 km
<10 km
< 15 km
N
N
%
N
%
N
%
N
%
PROXIMITY TO PM10 MONITORS
Atlanta
448411
3745
0.84
55720
12.43
128942
28.76
191136
42.63
Birmingham
147103
5682
3.86
44522
30.27
66145
44.96
77685
52.81
Boston
392413
3581
0.91
53764
13.70
126432
32.22
170493
43.45
Chicago
1094372
9242
0.84
115912
10.59
333566
30.48
560160
51.19
Denver
189663
3342
1.76
55670
29.35
121772
64.20
161144
84.96
Detroit
514431
3228
0.63
60935
11.85
136051
26.45
223387
43.42
Houston
683428
7244
1.06
154230
22.57
378960
55.45
484010
70.82
Los Angeles
2142800
9433
0.44
281562
13.14
929477
43.38
1654985
77.23
New York
n/a
rJa
rJa
n la
n la
n/a
n/a
n la
rJa
Philadelphia
657596
4484
0.68
51088
171
113232
17.22
215062
32.70
Phoenix
371104
19813
5.34
176270
47.50
288298
77.69
341603
92.05
Pittsburgh
255621
9344
3.66
77798
30.43
130064
50.88
163157
63.83
Riverside
488028
11292
2.31
150452
30.83
340809
69.83
394402
80.82
Seattle
215825
487
0.23
25714
11.91
63627
29.48
88950
41.21
St. Louis
296508
5359
1.81
63694
21.48
128961
43.49
162680
54.87
PROXIMITY TO PM2.S MONITORS
Atlanta
448411
1687
0.38
44678
9.96
177700
39.63
261042
58.21
Birmingham
147103
1785
1.21
30719
20.88
80301
54.59
101233
68.82
Boston
392413
14787
3.77
155069
39.52
239226
60.96
302977
77.21
Chicago
1094372
32528
2.97
491859
44.94
869986
79.50
996601
91.07
Denver
189663
3365
1.77
54486
28.73
135353
71.37
168509
88.85
Detroit
514431
8814
1.71
159732
31.05
334534
65.03
410145
79.73
Houston
683428
2952
0.43
55174
8.07
216596
31.69
319934
46.81
Los Angeles
2142800
24354
1.14
507262
23.67
1431687
66.81
1957788
91.37
New York
2597288
128534
4.95
1269125
48.86
1954534
75.25
2251450
86.68
Philadelphia
657596
18770
2.85
321148
48.84
481617
73.24
544596
82.82
Phoenix
371104
7397
1.99
89820
24.20
179137
48.27
220683
59.47
Pittsburgh
255621
6458
2.53
67986
26.60
146525
57.32
192872
75.45
Riverside
488028
8415
1.72
128696
26.37
275673
56.49
322315
66.04
Seattle
215825
1357
0.63
24146
11.19
65529
30.36
95466
44.23
St. Louis
296508
6077
2.05
77103
26.00
167662
56.55
202379
68.25
December 2008
3-163
DRAFT-DO NOT QUOTE OR CITE
-------
Table 3-31. Proximity to PM10 and PM2.5 monitors for adults older than 25 with at least a high
school degree by city. Percentages are given with respect to the total population per
city provided.
Region
Total CSA/CBSA
<1 km
<5 km
<10 km
< 15 km
N
N
%
N
%
N
%
N
%
PROXIMITY TO PM10 MONITORS
Atlanta
1417980
6269
0.44
102003
7.19
285569
20.14
508191
35.84
Birmingham
387984
8626
2.22
95125
24.52
171953
44.32
227554
58.65
Boston
1466291
6156
0.42
108181
7.38
295661
20.16
476645
32.51
Chicago
3024059
20451
0.68
321392
10.63
860819
28.47
1317793
43.58
Denver
733177
6235
0.85
91236
12.44
279592
38.13
468452
63.89
Detroit
1722217
4078
0.24
116031
6.74
283436
16.46
496870
28.85
Houston
1441713
9252
0.64
202238
14.03
545578
37.84
790301
54.82
Los Angeles
3531165
15995
0.45
367810
10.42
1291240
36.57
2405244
68.11
New York
n/a
rJa
rJa
n la
n la
n/a
n/a
n la
rJa
Philadelphia
2046124
8425
0.41
143098
6.99
412022
20.14
810010
39.59
Phoenix
1167897
23332
2.00
345572
29.59
780607
66.84
1045509
89.52
Pittsburgh
1041150
26717
2.57
281423
27.03
528748
50.78
690759
66.35
Riverside
1118856
15916
1.42
237268
21.21
660137
59.00
837766
74.88
Seattle
1142988
1680
0.15
77015
6.74
244983
21.43
403001
35.26
St. Louis
1023482
10048
0.98
142311
13.90
331032
32.34
453422
44.30
PROXIMITY TO PM2.S MONITORS
Atlanta
1417980
5392
0.38
146362
10.32
523634
36.93
855616
60.34
Birmingham
387984
4234
1.09
84724
21.84
235143
60.61
299384
77.16
Boston
1466291
27628
1.88
335506
22.88
675663
46.08
970310
66.17
Chicago
3024059
51344
1.70
976946
32.31
2008665
66.42
2540602
84.01
Denver
733177
4505
0.61
110496
15.07
365710
49.88
556030
75.84
Detroit
1722217
15101
0.88
364242
21.15
901878
52.37
1210638
70.30
Houston
1441713
2572
0.18
44544
3.09
207574
14.40
410350
28.46
Los Angeles
3531165
29011
0.82
660640
18.71
1979473
56.06
2875097
81.42
New York
5915247
184470
3.12
2034067
34.39
3596554
60.80
4409699
74.55
Philadelphia
2046124
34271
1.67
651790
31.85
1248293
61.01
1542036
75.36
Phoenix
1167897
8462
0.72
122557
10.49
296712
25.41
497653
42.61
Pittsburgh
1041150
18789
1.80
244276
23.46
578962
55.61
807780
77.59
Riverside
1118856
10547
0.94
184749
16.51
506166
45.24
658570
58.86
Seattle
1142988
5444
0.48
108082
9.46
343358
30.04
548899
48.02
St. Louis
1023482
13983
1.37
207288
20.25
493223
48.19
650270
63.54
December 2008
3-164
DRAFT-DO NOT QUOTE OR CITE
-------
Table 3-32. Proximity to PM10 and PM2.5 monitors for adults older than 25 with at least a college
degree by city. Percentages are given with respect to the total population per city
provided.
Region
Total CSA/CBSA
<1 km
<5 km
<10 km
< 15 km
N
N
%
N
%
N
%
N
%
PROXIMITY TO PM10 MONITORS
Atlanta
853073
6500
0.76
99524
11.67
250152
29.32
392639
46.03
Birmingham
157420
1344
0.85
24587
15.62
70055
44.50
108070
68.65
Boston
1090449
3003
0.28
132943
12.19
304776
27.95
447975
41.08
Chicago
1678748
4649
0.28
85800
5.11
269061
16.03
447794
26.67
Denver
479097
4347
0.91
56010
11.69
148519
31.00
257297
53.70
Detroit
676467
763
0.11
20620
3.05
59301
8.77
108423
16.03
Houston
761389
2340
0.31
89595
11.77
287098
37.71
434060
57.01
Los Angeles
2019245
6311
0.31
139325
6.90
566194
28.04
1150060
56.95
New York
n/a
n/a
n/a
n/a
n/a
n la
n/a
n la
rJa
Philadelphia
1037440
1629
0.16
46973
4.53
190538
18.37
459073
44.25
Phoenix
515058
6626
1.29
134934
26.20
324835
63.07
447886
86.96
Pittsburgh
396981
6281
1.58
105147
26.49
215427
54.27
292152
73.59
Riverside
312236
3870
1.24
53356
17.09
180833
57.92
241193
77.25
Seattle
660973
544
0.08
39609
5.99
155878
23.58
295547
44.71
St. Louis
435940
1950
0.45
35179
8.07
105568
24.22
171505
39.34
PROXIMITY TO PM2.S MONITORS
Atlanta
853073
5028
0.59
118328
13.87
382938
44.89
601057
70.46
Birmingham
157420
1848
1.17
36864
23.42
109871
69.79
140664
89.36
Boston
1090449
29378
2.69
248873
22.82
496229
45.51
708446
64.97
Chicago
1678748
15582
0.93
397700
23.69
1079947
64.33
1434516
85.45
Denver
479097
2566
0.54
73886
15.42
228579
47.71
356830
74.48
Detroit
676467
3403
0.50
105508
15.60
310644
45.92
451474
66.74
Houston
761389
436
0.06
6395
0.84
32141
4.22
110252
14.48
Los Angeles
2019245
12045
0.60
302402
14.98
950488
47.07
1424480
70.55
New York
3700528
157911
4.27
1364013
36.86
2223755
60.09
2767157
74.78
Philadelphia
1037440
21130
2.04
216512
20.87
537396
51.80
755892
72.86
Phoenix
515058
1977
0.38
32808
6.37
84812
16.47
191614
37.20
Pittsburgh
396981
3564
0.90
92830
23.38
214879
54.13
325673
82.04
Riverside
312236
1613
0.52
36761
11.77
136016
43.56
191453
61.32
Seattle
660973
1877
0.28
41826
6.33
178144
26.95
332915
50.37
St. Louis
435940
3719
0.85
83327
19.11
210846
48.37
288173
66.10
December 2008
3-165
DRAFT-DO NOT QUOTE OR CITE
-------
i Kilometers
0 5 10
20
30
40
r>
0 1f
30
60
90
i Kilometers
120
2000 Population Below Poverty
Line per Sq Mile
| 0 -204
| 205 -408
409 - 2038
2039 -4075
4076 - 10188
| 10189 -40750
j \ PM 2 5 Monitors (5 km buffer)
Figure 3-81. PM2.5 sampler density compared with numbers in poverty per square mile in the
Philadelphia CSA.
December 2008
3-166
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
3.8. Summary and Conclusions
3.8.1. Concentrations and Sources of Atmospheric PM
3.8.1.1. Ambient PM Variability and Correlations
Advances in understanding the spatiotemporal distribution of PM mass and constituents have
recently been made, particularly with regard to PM2 5 mass and chemical composition and ultrafine
concentrations. Emphasis in this ISA was on the period from 2005-2007 so that the most recent validated
EPA Air Quality System (AQS) data were used. Note, however, that a majority of U.S. counties were not
represented in AQS data, since their population densities fell below the regulatory monitoring threshold
for PM. Moreover, monitors reporting to AQS were not uniformly distributed across the U.S. or within
counties, and conclusions drawn from AQS data may not apply equally to all parts of a geographic region.
Furthermore, biases can exist for some PM constituents (and hence total mass) owing to volatilization
losses of nitrates and other semi-volatile compounds, and, conversely, to retention of particle-bound water
by hygroscopic species. The degree of spatial variability in PM was likely to be region-specific and
strongly influenced by region-specific sources and meteorological and topographic conditions.
Spatial Variability across the U.S.
County-scale, 24-h average concentration data for PMi0 and PM2 5 for 2005—2007 showed
considerable variability across the U.S. Figures 3-6 and 3-7 show county-scale coverage and average
concentrations for PMi0 and PM25. For PMi0, the highest reported annual average concentrations
(>51 (.ig/ni3) occurred in two counties in southern California and five counties in southern Arizona and
central New Mexico. The lowest reported annual average PMi0 concentrations (<20 (.ig/ni3) were within
114 counties distributed fairly uniformly across the U.S. For PM25, the highest reported annual average
concentration (>20 (.ig/ni3) were reported for six counties within the San Joaquin Valley and inland
southern California, as well as Jefferson County, AL (containing Birmingham) and Allegheny County, PA
(containing Pittsburgh). The lowest reported annual average PM2 5 concentrations (< 20 (.ig/ni3) were
contained within 237 counties distributed throughout the west, northeast, Florida and the Carolinas.
The concentration of PM25 relative to that of PM10 varied substantially by location, with a larger
fraction of PM mass in the coarse mode in cities with dryer climates (e.g., Phoenix and Denver) and a
larger fraction in the fine mode in eastern U.S. cities (e.g., Pittsburgh and Philadelphia). Limiting the
differential calculation of PM10.2.5 to low volume federal reference method (FRM) PM10 and PM2 5
monitors helps reduce sampling artifacts resulting from subtracting two independent mass measurements.
December 2008
3-167
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
However, this results in poor geographic coverage since few sites have the appropriate co-located
monitors for computing this difference. Figure 3-8 contains all U.S. counties where co-located low
volume FRM data was available for this calculation. Although the general understanding of PM
differential settling leads to an expectation of greater spatial heterogeneity in the PMi0.2.5 fraction,
deposition of particles as a function of size depends strongly on local meteorological conditions. Current
data coverage is insufficient to draw any meaningful conclusions regarding the spatial distribution of
PMio-2.5-
Spatial variability in PM2 5 components obtained from the Chemical Speciation Network (CSN)
varied considerably by species, including OC, EC, S042 , N03 and NH/ (see Section 3.5.1.1). The
highest annual average OC concentrations (>5 (.ig/nr1) were observed in the western and southeastern U.S.
Concentrations in the western U.S. peaked in the fall and winter, while concentrations in the Southeast
peaked anytime between spring and fall. EC exhibited less seasonality than OC and was particularly
stable in the eastern half of the U.S. Annual average EC concentrations greater than 1.5 (ig/m3 were
present in Los Angeles, Pittsburgh, New York and El Paso. Concentrations of S042 were higher in the
eastern U.S. as a result of higher S02 emissions in the East, compared with the West. There is also
considerable seasonal variability with higher S042 concentrations in the summer months when the
oxidation of S02 proceeds at a faster rate than during the winter. N03 concentrations were highest in
California, with annual averages >4 (ig/m3 at many monitoring locations. There were also elevated levels
ofMV in the Upper Midwest (>2 (ig/m3), particularly in the winter. In general, N03 was higher in the
winter across the country, in part as a result of temperature-driven partitioning and volatilization.
Exceptions existed in Los Angeles and Riverside, where high N03 readings appeared year round.
Concentrations ofNH4+ were similar to concentrations of N03 or S042 throughout the U.S. Clearly,
there is variation in both PM2 5 mass and composition by city, some of which might be due to regional
differences; however, there are too many controlling variables (e.g. meteorology, sources, topography)
which are varied and too poorly characterized at this scale to allow conclusions to be drawn regarding
PM2 5 composition across all cities within a given geographic region.
Variability in PM2 5 components across the U.S. was examined by focusing on fifteen metropolitan
areas chosen based on their geographic distribution and coverage in recent health effects studies (see
Section 3.5.1.1). The urban areas selected were Atlanta, Birmingham, Boston, Chicago, Denver, Detroit,
Houston, Los Angeles, New York, Philadelphia, Phoenix, Pittsburgh, Riverside, Seattle and St. Louis. On
an annual average basis, sulfate was the dominant PM2 5 component in the eastern cities, ranging from
42% of PM2 5 mass in Chicago to 56% in Pittsburgh. Organic carbon mass (OCM) was the next largest
component. In the western cities, OCM was the largest constituent of PM2 5 on an annual basis, ranging
from 34% in Los Angeles to 58% in Seattle. Sulfate, nitrate and crustal material were all important
December 2008
3-168
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
components in the western cities analyzed. Sulfate ranged from 18% in Denver to 32% in Los Angeles.
Nitrate was particularly large in Riverside (22%), Los Angeles (19%) and Denver (15%); crustal material
constituted a substantial fraction of PM25 year-round in Phoenix (28%) and Denver (16%), and during the
summer in Houston (26%), even though the annual average was much lower (11%).
Spatial Variability on the Urban and Neighborhood Scales
In general, PMi0 has a shorter atmospheric lifetime than PM2 5 because PMi0 contains larger
particles which have higher settling velocity. As a result, local emission sources often dominate PMi0
annual average mass concentrations at particular monitors, while PM2 5 mass concentrations are more
homogeneously distributed (see Section 3.5.1.2). Therefore, as an example, using the 15 cities listed
above, there was considerably less decline in the correlation between monitors as a function of distance
for PM2 5 than for PMi0. Furthermore, correlations between PM2 5 concentrations exhibited substantially
less scatter. For PMi0, Atlanta, Boston, Denver, Los Angeles, New York City, Philadelphia, Phoenix,
Pittsburgh and Riverside all showed relatively high correlations as a function of distance (average
correlation of 0.75 at 40 km or greater monitor separation), while Birmingham, Chicago, Detroit, Houston
and St. Louis had correlations that dropped off much more quickly with distance (average correlation of
0.75 at 6 km or less monitor separation). The Seattle data only included two PMi0 monitoring sites, thus
providing insufficient information to draw any conclusions. For PM2 5, most metropolitan areas exhibited
high correlations (generally >0.75) out to a distance of 100 km. Notable exceptions were Denver, Los
Angeles and Riverside, where correlations dropped below 0.75 somewhere between 20 and 50 km.
Insufficient data were available in the 15 metropolitan areas to perform similar analyses for PM10.2 5 using
co-located, low volume FRM monitors.
Population density and associated building density are important determinants of the spatial
distribution of PM concentrations. Inter-sampler correlations as a function of distance between monitors
obtained for sampler pairs located less than 4 km apart (i.e., on a neighborhood scale) showed a shallower
slope for PM2 5 than for PMi0. Average correlation was maintained at 0.93 for PM2 5, while it dropped to
0.70 for PM10 (see Section 3.5.1.3).
Few studies performed direct comparisons of ultrafine particle measurements at multiple locations
within an urban area. A decrease in the number of ultrafine particles was demonstrated with shifts from a
dominant mode at around 10 nm within 20 m of a freeway to a flattened dominant mode at around 50 nm
at a distance of roughly 100-150 m. At the same time, accumulation mode particle number concentration
remained relatively constant to within -300 m from the freeway. These findings suggest a high degree of
spatial heterogeneity in ultrafine particles compared with accumulation mode particles on the urban scale.
December 2008
3-169
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
3.8.1.2. Temporal Variability
Trends in PM10 concentrations show a steady decline from 1988 to 2007 in all 10 EPA Regions. A
steady decrease in PM2 5 concentrations from 1999 (the beginning of nationwide monitoring for PM2 5) to
2007 was observed in all 10 EPA Regions, with the three-year average of the 98th percentile of 24-h
PM2 5 concentrations dropping 10% over this time period.
Using hourly PM observations in the 15 metropolitan areas, diel variation showed peaks that differ
by pollutant and region. For PMi0, all areas showed a gradual morning increase in mean concentrations
starting at approximately 6:00 am on weekdays, corresponding with both the start of morning rush hour
and break-up of the overnight inversion layer. The magnitude and duration of this peak varied
considerably by metropolitan area. For PM2 5, a similar morning peak was observed starting at
approximately 6:00 am in all cities except Pittsburgh, where elevated overnight PM2 5 obscures any
morning peak. There was also an evening PM2 5 concentration peak that was broader than the morning
peak and extended into the overnight period, reflecting the concentration increase caused by the usual
collapse of the mixed layer after sundown (see Section 3.5.2.3).
Studies indicate that ultrafine particles in urban environments exhibit similar two-peaked diel
patterns in Los Angeles and the San Joaquin Valley as well as in Kawasaki City, Japan and Copenhagen,
Denmark (see Section 3.5.2.3). The afternoon peak in ultrafine particles likely represents the combination
of primary source emissions such as evening rush hour traffic and photochemical formation of secondary
organic aerosol.
3.8.1.3. Correlations between Copollutants
Correlations between PM and gaseous copollutants including S02, N02, carbon monoxide (CO)
and 03 varied both seasonally and spatially between and within metropolitan areas. On average, PM10 and
PM2 5 were correlated with each other better than with the gaseous copollutants. There was relatively little
seasonal variability in the mean correlation between PM in both size fractions and S02 and N02. CO,
however, showed higher correlations with PM10 and PM2 5 on average in the winter compared with the
other seasons. This seasonality results in part because a larger fraction of PM is primary in origin during
the winter. To the extent that this primary component of PM is associated with common sources of N02
and CO, then higher correlations with these gaseous co-pollutants are to be expected. Increased
atmospheric stability in colder months would also reinforce these associations.
The correlation between daily maximum 8-h average 03 and PM showed the highest degree of
seasonal variability with positive correlations on average in the spring, summer and fall, and negative
correlations on average in the winter. This situation arises as the result of seasonal differences in sources
December 2008
3-170
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
and photochemical production of secondary PM25 and 03. However, this relationship is not found in all
cities examined (e.g., Birmingham, Boston and St. Louis).
3.8.1.4. Measurement Techniques
Reliable methods have been developed to measure real-time PM mass concentrations (e.g., FDMS-
TEOM). Real-time (or continuous and semi-continuous) measurement techniques are also available for
PM species, such as PILS for multiple ions analysis and AMS for multiple components analysis.
Advances have also been achieved in PM organic speciation (e.g. TD-GC-MS). (For addition information
see Section 3.4.)
3.8.1.5. PM Source Characteristics
PM in the atmosphere contains both primary (i.e., emitted directly by sources) and secondary
components, which can be anthropogenic or natural in origin. Secondary components are produced by the
oxidation of precursor gases such as S02, NOx and ammonia (NH3) and organic compounds. The largest
sources of primary PM2 5 on a nationwide basis are wildfires, road dust, and electricity-generating units
(EGUs), with road dust being the largest single source of PMi0 according to the National Emissions
Inventory (NEI).
Developments in the chemistry of formation of secondary organic aerosol (SOA) indicate that
oligomers are likely a major component of OC in aerosol samples. Until a few years ago, the oxidation of
terpenes and aromatic compounds were considered as sources of SOA, but not the oxidation of isoprene.
However, recent observations suggest that small, but important quantities of SOA are formed from
isoprene oxidation. Gasoline engines have been found to emit a mix of OC, EC, and nucleation-mode
heavy and large polycyclic aromatic hydrocarbons on which unspent fuel and trace metals condense,
while diesel particles are composed of a soot nucleus on which S042 and hydrocarbons condense.
Current inventories of emissions from combustion sources overestimate the primary component of
organic aerosol and underestimate the semi-volatile components in the emissions. This situation results
from the lack of capture of evaporated semi-volatile components upon dilution in standard emissions
tests. As a result, near-traffic sources of organic aerosol are underestimated, however, farther downwind
the overall formation rate of SOA increases as a result of the oxidation of these semi-volatile components.
3.8.1.6. Source Contributions to PM
Results of receptor modeling calculations indicate that PM2 5 is produced mainly by combustion of
fossil fuel, either by stationary sources or by transportation. It is apparent that a relatively small number
December 2008
3-171
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
of source categories, compared to the total number of chemical species that typically are measured in
ambient monitoring source receptor model studies, are needed to account for the majority of the observed
mass of PM in these studies. A compilation of study results shows that secondary sulfate (mainly from
EGUs), nitrate (from the oxidation of NOx emitted mainly from transportation and EGUs), and primary
mobile source categories constitute most of PM2.5 (and PMi0) in the East. Fugitive dust, found mainly in
the PM10.2 5 size range, represents the largest source of ambient PMi0 in many locations in the western
U.S. Quoted uncertainties in the source apportionment of constituents in ambient aerosol samples
typically range from 10 to 50%. An intercomparison of source apportionment techniques indicated that
the same major source categories of PM2 5 were consistently identified by several independent groups
working with the same data sets. Soil-, sulfate-, residual oil-, and salt-associated mass were most clearly
identified by the groups. Other sources with more ambiguous signatures, such as vegetative burning and
traffic-related emissions were less consistently identified.
Spatial variability in source contributions across urban areas is an important consideration in
assessing the likelihood of exposure error in epidemiologic studies relating health endpoints to sources.
Concepts similar to those for using ambient concentrations as surrogates for personal exposures apply
here. Studies for PM2.5 indicate that intra-urban variability increases in the following order: regional (e.g.
secondary S042 from EGUs) < area (e.g. on road mobile sources) < point (e.g. stacks) sources. Only one
study was available for PM10.2 5, indicating a similar ordering, but without a regional component (resulting
from the short lifetime of PM10.2.5 compared to transport times on the regional scale).
3.8.1.7. Policy-Relevant Background
The background concentration of PM that is useful for risk and policy assessments informing
decisions about the NAAQS are referred to as policy-relevant background (PRB) concentrations. PRB
concentrations are those concentrations that would occur in the U.S. in the absence of anthropogenic
emissions in continental North America (defined here at the U.S., Canada and Mexico). PRB
concentrations include contributions from natural sources everywhere in the world and from
anthropogenic sources outside these three countries. Background levels so defined facilitate separation of
pollution levels that can be controlled by U.S. regulations (or through international agreements with
neighboring countries) from levels that are generally uncontrollable by the U.S. Contributions to policy-
relevant background (PRB) levels of PM include both primary and secondary natural and anthropogenic
components (see Section 3.6). PRB concentrations for the continental U.S. were estimated using a
deterministic, continental scale chemistry-transport model (CTM) using results from the GEOS-Chem
global scale CTM as boundary conditions. PRB concentrations of PM2 5 were estimated to be less than 1
(ig/m3 on an annual basis and maximum daily average values generally range from about 3 to 20 (ig/m3
December 2008
3-172
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
with a peak as high 63 (ig/m3 as at the nine national park sites across the U.S. that were used for model
evaluation.
3.8.2. Human Exposure
This section summarizes the findings from the recent exposure assessment literature, which include
the assessment of exposure to ambient PM, infiltration of ambient PM to indoor environments, and source
apportionment of exposure. This summary is intended to support the interpretation of the findings from
epidemiologic studies. For a more detailed explanation see Section 3.7.
3.8.2.1. Outdoor Exposure to Ambient PM
The correlation between the PM concentration measured at a central community ambient monitor
and the true community average concentration depends on the spatial distribution of the PM, selection of
the monitoring site chosen to represent the community average, and division of the community by terrain
features or source locations into several sub-communities that differ in the temporal pattern of pollution.
Some studies, conducted mainly in Europe, have found that personal PM2 5 and PMi0 exposures for
pedestrians in street canyons could be much higher than ambient concentrations measured by urban
background ambient monitors. As a result, ambient monitors located at background, central urban, road
side, or near-residential sites might not reflect peak exposures to some individuals in a community.
Ambient monitor height also affects estimates of exposure because PM concentration varies as a function
of height. Within a street canyon, changes in wind direction and speed cause significant variability over a
small distance, with findings showing up to a two order of magnitude change in benzo[a]pyrene
concentrations across a street canyon. Wind tunnel studies have shown street canyon effects exist for
suburban and not just for downtown, heavily urbanized settings.
3.8.2.2. Indoor and Personal Exposure to Ambient PM
PM infiltration factors, Finf, depend on particle size, chemical composition, season, and region of
the country. Infiltration can best be modeled dynamically based on a distribution of air exchange and
deposition PM loss rates rather than being represented by a single value. There is significant variability
within and across regions of the country with respect to indoor exposures to ambient PM. Infiltrated
ambient PM concentrations depend in part on the ventilation properties of the building or vehicle in which
the person is exposed. Season is important to PM infiltration because it affects the ventilation practices
used, and ambient temperature and humidity conditions affect the transport, dispersion, and size
December 2008
3-173
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
distribution of PM. Residential air exchange rates have been observed to be higher in summer for regions
with low air conditioning usage, and regional differences in air exchange rates (Southwest < Southeast
< Northeast < Northwest) also reflect ventilation practices. Differential infiltration occurs as a function of
PM size and composition. PM infiltration is largest for accumulation mode particles, and decreases for
ultrafine PM lost to diffusion and for coarse particles lost through inertial impaction mechanisms.
Infiltration is also affected by variations in particle composition and volatility. For example, EC or black
carbon (BC) infiltrates more readily than OC. Differential infiltration can affect both exposure estimates
and PM toxicity.
Emission inventories and source apportionment studies suggest that sources of PM exposure vary
by region. Comparison of studies performed in the eastern U.S. with studies performed in the western
U.S. suggest that the contribution of S042 to personal exposure is higher for the East (16-46%) compared
with the West (-4%) and that motor vehicle emissions and secondary N03 are larger sources of personal
exposure for the West (-9%) as compared with the East (-4%). Results of source apportionment studies
of personal exposure to S042 indicate that personal S042 exposures are mainly attributable to ambient
sources. Source apportionment for OC and EC is difficult because they originate from both indoor and
outdoor sources. Exposure to OC of indoor and outdoor origin can be distinguished by the presence of
aliphatic C-H groups generated indoors, since outdoor concentrations of aliphatic C-H are low. Trace
metal studies have shown variable results regarding personal exposure to ambient constituents with
significant variation among cities and over seasons that can be related to incinerator operation, fossil fuel
combustion, biomass combustion (wildfires), and presence of crustal materials in the built environment,
among other sources.
3.8.2.3. Implications for Epidemiologic Studies
Variations in PM and its components could lead to errors in using ambient PM measures as
surrogates for exposures to PM. PM2 5 and PMi0 concentrations are relatively well-correlated across
monitors in the urban areas examined. Correlation coefficients tend to be lower, and concentration
differences tend to be higher between PMi0 monitoring sites than between PM2 5 monitoring sites. Even if
PM2 5 and PMi0 concentrations measured at sites within an urban area are highly correlated, significant
differences in their concentrations can occur on any given day. The degree of spatial uniformity in PMi0
and PM2 5 concentrations in urban areas varies across the country. Current information suggests that
PMio-2.5 and some PM components are more spatially variable than PM2 5. These factors should be
considered in using data obtained from monitoring networks to estimate community-scale human
exposure to ambient PM, and caution should be exercised in extrapolating conclusions obtained from one
urban area to another.
December 2008
3-174
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Community, time-series epidemiologic studies use the average community PM concentration as a
surrogate for the average personal exposure to ambient PM. The resulting health effect risk estimate,
based on the average community ambient concentration, differs from the risk that would be estimated if
the average community ambient exposure were used in the epidemiologic study. This difference is given
by the average ambient exposure factor. However, the risk estimate based on the ambient concentration
gives the change in health effects resulting from a change in ambient concentration of PM and is,
therefore, an appropriate measure for risk assessment and risk management. Variations in ambient
concentrations across a community, variations in individual ambient exposures around the community
average, and seasonal or daily variation in the ambient exposure factor may increase standard errors of
PM health effects estimates, making it more difficult to detect a true underlying association between the
correct exposure metric and the health outcome studied. The use of the community average ambient PM
concentration as a surrogate for the community average personal exposure to ambient PM is not expected
to change the principal conclusions from PM epidemiologic studies that use community average health
and pollution data (U.S. EPA, 2004). Several recent studies support this by showing how the ambient
component of personal exposure to PM2 5 could be estimated using various tracer and source
apportionment techniques and that is highly correlated with ambient concentrations of PM25. These
studies also show that the non-ambient component of personal exposure to PM2 5 is basically uncorrelated
with ambient PM2 5 concentrations. For long-term studies that use differences in long-term community
average ambient PM concentrations as an exposure metric, the effect of possible community-to-
community differences in the average ambient exposure factor or in the average non-ambient exposure
are less understood. For panel epidemiologic studies, the most appropriate exposure metric may depend
on the health outcome measured. However, sufficient information should be obtained to enable
determining the association of the health outcome with ambient concentration, ambient exposure, non-
ambient exposure, and total personal exposure.
A number of studies have examined whether gaseous copollutants could act as surrogates for
exposure to ambient PM. Several studies have concluded that ambient concentrations of 03, N02, and
S02 are associated with the ambient component of personal exposure to total PM2 5 as opposed to the
ambient component of personal exposures to the gases. However, in some studies this result may have
arisen in part because personal exposure to the gases was often beneath the detection limits of the
personal monitoring devices. Thus, the evidence that ambient gases can be considered surrogates of PM2.5
exposure is mixed. It is likely that associations between ambient gases and personal exposure to PM2 5 of
ambient origin exist, but they are complex and vary by season and location.
December 2008
3-175
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Chapter 4. Dosimetry
4.1. Introduction
Particle dosimetry refers to the characterization of deposition, translocation, clearance, and
retention of particles and their constituents within the respiratory tract and extrapulmonary tissues. This
chapter summarizes basic concepts presented in dosimetry chapters of the 1996 and 2004 PM AQCDs
(U.S. EPA, 1996, 2004), and updates the state of the science based upon new literature appearing since
publication of these PM AQCDS. Although our basic understanding of the mechanisms governing
deposition and clearance of inhaled particles has not changed, there has been significant additional
information on the role of certain biological determinants such as gender, age and lung disease on
deposition and clearance. Additionally, new studies have further characterized the retention and
translocation of ultrafine particles (also commonly referred to as nanoparticles) following deposition in
the respiratory tract.
The dose from inhaled particles deposited and retained in the respiratory tract is governed by a
number of factors. These include exposure concentration and duration, activity and ventilatory
parameters, and particle properties (e.g., particle size, hygroscopicity, and solubility in airway fluids and
cellular components). The basic characteristics of particles as they relate to deposition and retention, as
well as anatomical and physiological factors influencing particle deposition and retention, were discussed
in depth in Chapter 10 of 1996 PM AQCD and updated in Chapter 6 of the 2004 PM AQCD. Species
differences between humans and rats in particle exposures, deposition patterns, and pulmonary retention
were also reviewed in Brown et al. (2005). The current review of PM dosimetry focuses mainly on issues
that may affect the susceptibility of an individual to adverse effects as well as issues that affect our ability
to extrapolate findings between studies (e.g., in vitro to in vivo) and between species.
4.1.1. Size Characterization of Inhaled Particles
Particle size is a major determinant of the fraction of inhaled particles depositing in and cleared
from various regions of the respiratory tract. The distribution of particle sizes in an aerosol is typically
described by the lognormal distribution (i.e., the situation in which the logarithms of particle diameter are
distributed normally). The geometric mean is the median of the distribution, and the variability around the
median is the geometric standard deviation (GSD or og) and is given by:
December 2008
4-1
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
GSD = cj = ^84% = ^50%
^50% ^16%
Equation 4-1
where: di6o/o, d50o/o, d84o/o are the particle diameters associated with the 16th, 50th (i.e. the median), and the
84th percentiles from the cumulative frequency distribution of particle sizes. Note that the GSD is always
greater than one. The particle size associated with any percentile of the distribution, dl5 is given by:
d = d
-------
transition region between ciliated and fully alveolated airways, are found in humans, dogs, ferrets, cats,
and monkeys. Respiratory bronchioles are absent in rats and mice and abbreviated in hamsters, guinea
pigs, oxen, sheep, and pigs. The branching structure of the ciliated bronchi and bronchioles also differs
between species from being a rather symmetric and dichotomous branching network of airways in
humans to a more monopodial branching network in other mammals.
Posterior
Nasal Passage
Nasal Part
Oral Part
ET.
Extrathoracic
Region
Pharynx
Larynx
Trachea
Main Bronchi
Tracheobronchial
Region
BB
Bronchi
Bronchioles
bb
Bronchiolar Region
Alveolar Interstitial
AI
Ai.
Bronchioles
— Terminal Bronchioles
Respiratory Bronchioles
Alveolar
Region
AI
Alveolar Duct +
Alveoli
Source: Based on ICRP (1994).
Figure 4-1. Diagrammatic representation of respiratory tract regions in humans. Structures are
anterior nasal passages, ETi; oral airway and posterior nasal passages, ET2; bronchial
airways, BB; bronchioles, bb; and alveolar interstitial, AI.
December 2008
4-3
DRAFT-DO NOT QUOTE OR CITE
-------
1 Another species difference relevant to particle dosimetry is the route of breathing. For instance, rats
2 are obligate nose breathers, whereas most humans are oronasal breathers who breathe through the nose
3 when at rest and increasingly through the mouth with increasing activity level. There is inter-individual
4 variability in the route by which people breathe. Most people, 87% (26 of 30) in the Niinimaa et al.
5 (1981) study, breathed through their nose until an activity level was reached where they switched to oro-
6 nasal breathing. Thirteen percent (4 of 30) of the subjects, however, were oronasal breathers even at rest.
7 These two subject groups are commonly referred to in the literature (e.g., (ICRP, 1994) as ""normal aug-
8 menters" and "mouth breathers," respectively. In contrast to healthy subjects, Chadha et al. (1987) found
9 that the majority (11 of 12) of patients with asthma or allergic rhinitis breathe oronasally even at rest.
Bronchus
Bronchiolus
Aveolus
b.
Air
Air
Illllllll
Liquid
iiiiiiiii
Tissue
Tissue
Tissue
Tissue
Air
Source: Panel (a) reproduced with permission (ER Weibel, Design and structure of the human lung,
In: Pulmonary Diseases and Disorders, ed. AP Fishman, McGraw-Hill, New York, 1980, p. 231)(Fishman and Elias, 1980).
Figure 4-2. Structure of lower airways with progression from the large airways to the alveolus.
Panel (a) illustrates basic airway anatomy. Structures are epithelial cells, EP;
basement membrane, BM; smooth muscle cells, SM; and fibrocartilaginous coat, FC.
Panel (b) illustrates the relative amounts of liquid, tissue, and blood with distal
progression.
December 2008
4-4
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
The site of particle deposition within the respiratory tract has implications related to lung retention
and surface dose of particles as well as potential systemic distribution of particles or their constituents.
Figure 4-2 illustrates the progressive change in airway anatomy with distal progression into the lower
respiratory tract. In the bronchi there is a thick liquid lining and mucociliary clearance rapidly moves
deposited particles toward the mouth. In general, in the bronchi, only highly soluble materials moving
from the air into the liquid layer will have systemic access via the blood. With distal progression, the
protective liquid lining diminishes and clearance rates slow. Soluble compounds and some poorly soluble
ultrafine particles may cross the air-liquid interface to enter the tissues and the blood especially in the
alveolar region.
4.2. Particle Deposition
Inhaled particles may be either exhaled or deposited in the extrathoracic (ET), tracheobronchial
(TB), or alveolar (A) region. A particle becomes deposited when it moves from the airway lumen to the
wall of an airway. The deposition of particles in the respiratory tract depends primarily on inhaled particle
size, route of breathing (through the nasal or oronasal), tidal volume (Vt), breathing frequency (/), and
respiratory tract morphology. The distinction between air passing through the nose versus the mouth is
important since the nasal passages more effectively remove inhaled particulate than the oral passage.
Respiratory tract morphology, which affects particle transport and deposition, varies between species, the
size of an animal or human, and health status.
The fraction of inhaled aerosol becoming deposited in the human respiratory tract has been
measured experimentally. Studies, using light scattering or particle counting techniques to quantify the
amount of aerosol in inspired and expired breaths, have characterized total particle deposition for varied
breathing conditions and particle sizes. The vast majority of in vivo data on the regional particle
deposition has been obtained by scintigraphic methods. These data have shown highly variable regional
deposition with sites of highly localized deposition or "hot spots" in the obstructed lung relative to the
healthy lung. Even in the healthy lung, "hot spots" occur in the region of airway bifurcations.
Mathematical models aid in predicting the mixed effects of particle size, breathing conditions, and lung
volume on total and regional deposition.
In order to potentially become deposited in the respiratory tract, particles must first be inhaled. The
inspirable particulate mass fraction of an aerosol is that fraction of the ambient airborne particles that can
enter the uppermost respiratory tract compartment, the head (Soderholm, 1985). The American
Conference of Governmental Industrial Hygienists (ACGIH) and the International Commission on
Radiological Protection (ICRP) have established inhalability criteria for humans (ACGIH, 2005; ICRP,
December 2008
4-5
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
1994). These criteria are indifferent to route of breathing and assume random orientation with respect to
wind direction. They are based on experimental inhalability data for dae < 100 |im at wind speeds of
between 1 and 8 m/s. For ACGIH criterion, inhalability is 97% for an dae = 1 (mi, 87% for an dae = 5 (mi,
77% for an dae = 10 (jm, and plateaus at 50% dae above -40 (mi. The ICRP criterion, which also plateaus
at 50% for very large dae, does not become of real importance until an dae = 5 (mi where inhalability is
97%. Dai et al. (2006) reported slightly lower nasal particle inhalability in humans during moderate
exercise than rest (e.g., 89.2 vs. 98.1% for 13 (mi particles, respectively). Nasal particle inhalability is
similar between an adult and 7-year-old child (Hsu and Swift, 1999). Inhalability into the mouth from
calm air in humans also becomes important for dae >10 |im (Anthony and Flynn, 2006; Brown, 2005).
Unlike the inhalability from high wind speeds which plateaus at 50% for dae greater than -40 (mi, particle
inhalability from calm air continues to decrease toward zero with increasing dae.
Inhalability data in laboratory animals, such as rats, are only available for breathing from relatively
calm air (velocity < 0.3 m/s). For nasal breathing, inhalability becomes an important consideration for dae
of above 1 |im in rodents and 10 |im in humans (Menache et al., 1995). The inhalability of particles
having dae of 2.5, 5, and 10 (j,m is 80, 65, and 44% in rats, respectively, whereas it only decreases to 96%
for a dae of 10 |im in humans during nasal breathing (Menache et al., 1995). Asgharian et al. (2003)
suggested that an even more rapid decrease in inhalability with increasing dae may occur in rats.
Inhalability is a particularly important consideration for rodent exposures to coarse particles. Section 4.2.3
provides additional discussion of interspecies patterns of particle deposition.
4.2.1. Mechanisms of Deposition
Particle deposition in the lung is predominantly governed by diffusion, impaction, and
sedimentation. Most discussion herein focuses on these three dominant mechanisms of deposition. Simple
interception, which is an important mechanism of fiber deposition, is not discussed in this chapter.
Electrostatic and thermophoretic forces as mechanisms of deposition have not been thoroughly evaluated
and receive limited discussion. Some generalizations with regard to deposition by these mechanisms
follows, but should not be viewed as absolute rules. Both experimental studies and mathematical models
have demonstrated that breathing patterns can dramatically alter regional and total deposition for all sized
particles. The combined processes of aerodynamic and diffusive (or thermodynamic) deposition are
important for particles in the range of 0.1 (mi to 1 (mi. Aerodynamic processes predominate above and
thermodynamic processes predominate below this range.
Diffusive deposition, by the process of Brownian diffusion, is the primary mechanism of deposition
for particles having physical diameters of less than 0.1 (mi. For particles having physical diameters of
December 2008
4-6
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
roughly between 0.05 and 0.1 (jm, diffusive deposition occurs mainly in the small distal bronchioles and
the pulmonary region of the lung. However, with further decreases in particle diameter below -0.05 jjxn,
increases in particle diffiisivity shift more deposition to the proximally to the bronchi and ET regions.
Governed by inertial or aerodynamic properties, impaction and sedimentation increase with dae.
When a particle has sufficient inertia, it is unable to follow changes in flow direction and strikes a surface
thus depositing by the process of impaction. Impaction occurs predominately at bifurcations in the
proximal airways, where linear velocities and secondary eddies are at their highest. Sedimentation, caused
by the gravitational settling of a particle, is most important in the distal airways and pulmonary region of
the lung. In these regions, residence time is the greatest and the distances that a particle must travel to
reach the wall of an airway are minimal.
The electrical charge on some particles may result in an enhanced deposition over what would be
expected from size alone. With an estimated charge of 10-50 negative ions per 0.5 |im particle, Scheuch et
al. (1990) found deposition in humans (VT = 500 ml,/= 15 min"1) to increase from 13.4% (no charge) to
17.8% (charged). This increase in deposition is thought to result from image charges induced on the
surface of the airway by charged particles. Yu (1985) estimated a charge threshold level above which
deposition fractions would be increased of about 12, 30, and 54 for 0.3, 0.6, and 1.0 ^m diameter
particles, respectively. Electrostatic deposition is generally considered negligible for particles below
0.01 (mi because so few of these particles carry a charge at Boltzmann equilibrium. This mechanism is
also thought to be a minor contributor to overall particle deposition, but it may be important in some
laboratory studies due to specific aerosol generation techniques such as nebulization. Laboratory methods
such as passage of the aerosols through a Kr-85 charge neutralizer prior to inhalation are commonly used
to mitigate this effect.
The National Radiological Protection Board (NRPB) recently evaluated the potential for corona
discharges from high voltage power lines to charge particles and enhance particulate doses (NRPB, 2004).
They concluded that electrostatic effects would be the most important for particles in the size range from
about 0.1-1 (mi, where deposition may theoretically increase by a factor of three to ten. However, given
that the small fraction of ambient particles would pass through the corona to become charged, the small
range of relevant particle sizes (0.1-1 |im). and the subsequent required transport of charged particles to
expose individuals; the NRPB concluded that effects, if any, of electric fields on particle deposition in the
human respiratory tract would likely be minimal.
Thermophoretic forces on particles occur due to temperature differences between respired air and
respiratory tract surfaces. Temperature gradients of around 20°C are thought to produce sufficient
thermophoretic force to oppose diffusive and electrostatic deposition during inspiration and to perhaps
augment deposition by these mechanisms during expiration (Jeffers, 2005). Thermophore sis is only
December 2008
4-7
DRAFT-DO NOT QUOTE OR CITE
-------
1 relevant in the extrathoracic and large bronchi airways and reduces to zero as the temperature gradient
2 decreases deeper in the lung. Theoretical analysis of thermophoresis has been done for smooth walled
3 tubes and is important over distances that are several orders of magnitude smaller than the diameter of the
4 trachea. The alteration of the flow patterns by airway surface features such as cartilaginous rings may
5 affect particle transport and deposition over far greater distances than thermophoretic force.
4.2.2. Deposition Patterns
6 Knowledge of sites where particles of different sizes deposit in the respiratory tract and the amount
7 of deposition therein is necessary for understanding and interpreting the health effects associated with
8 exposure to particles. Particles deposited in the various respiratory tract regions are subjected to large
9 differences in clearance mechanisms and pathways and, consequently, retention times. Deposition
10 patterns in the human respiratory tract were described in considerable detail in dosimetry chapters of prior
11 PM AQCD (U.S. EPA, 1996, 2004); as such, they are only briefly described here.
1.0
iotal y
nose
Breathing
Mouih
Breathing
Open Symbols MPPD Model
Closed Symbols ICRP Mode!
<$> and ^ Total
C
o
Total
2
IL
and []
0.6
I 0.4
I
ET
o.o
a
a
0.6
Mose
Breatning
Mouth
Breathing
0-5
U.
0.3
0.3
0.2
TB
TB
o.o
0.1
1
10
0.01
0.1
1
10
Diameter, |jm
Diameter, \im
Figure 4-3 Comparison of total and regional deposition results from the ICRP and the MPPD
models for a resting breathing pattern (Vt = 625 ml, f= 12 min-1). Panels a-b are for
nose breathing. Panels c-d are for mouth breathing.
December 2008
4-8
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Open Symbols MPPD Model
Closed Symbols ICRP Model
^ and ^ Total
O and # ET
~ and ¦ TB
Nose Breathing
Diameter, pm Diameter, pm
Figure 4-4 Comparison of total and regional deposition results from the ICRP and the MPPD
models for a light exercise breathing pattern (Vt = 1250 ml, f= 20 min-1). Panels a-b are
for nose breathing. Panels c-d are for mouth breathing.
Mouth Breathing
Predicted total and regional deposition for an adult male during rest and light exercise are
illustrated in Figures 4-3 and 4-4, respectively. Although these figures were provided in Chapter 6 of the
2004 PM AQCD, they are reproduced here to illustrate changes in deposition as a function of particle size
and breathing conditions. The predictions were based on two publicly available particle deposition
models, the ICRP (1994) and the MPPD model (Version 1.0, ©2002). The ICRP (1994) model was
implemented by LUDEP (Version 2.07, June 2000). The MPPD model was developed by the CUT
Centers for Health Research with support from the Dutch National Institute of Public Health and the
Environment. The MPPD model (Version 2) is currently available from The Hamner Institutes for Health
Sciences, RTP, NC.
4.2.2.1. Total Respiratory Tract Deposition
The efficiency of deposition in the respiratory tract may generally be described as a "U-shaped"
curve on a plot of deposition efficiency versus the of log particle diameter. Total deposition shows a
minimum for particle diameters in the range of 0.1 to 1.0 |im. where particles are small enough to have
minimal sedimentation or impaction and sufficiently large so as to have minimal diffusive deposition.
Total deposition does not decrease to zero for any sized particle because of mixing between particle laden
December 2008
4-9
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
tidal air and residual lung air. The particles mixed into residual air remain in the lung following a breath
and are removed on subsequent breaths or gradually deposited. Total deposition approaches 100% for
particles of roughly 0.01 |im (physical diameter) due to diffusive deposition and for particles of around
10 |im (aerodynamic diameter) due to the efficiency of sedimentation and impaction.
Total human respiratory tract deposition, as a function of particle size, is depicted in Figure 4-5.
These experimental data were obtained by using monodisperse spherical test particles in healthy adults
during controlled breathing on a mouthpiece. Despite the control of inhaled particle size and breathing
conditions, there is variability in deposition efficiencies due to inter-individual differences in lung size
and anatomical variability in airway dimensions and branching patterns.
£
O
o
ro
c
o
o
CL
a)
Q
Figure 4-5.
Particle Diameter (jjm)
Male Vt = 500 mi-
Female
Q = 250 mUs
* = p < 0.05
Source: Data from Kim and Hu (1998) and Kim and Jaques (2000).
Total lung deposition measured in healthy adults (ultrafine, 11 M, 11 F, 31 ± 4 years;
fine and coarse, 11 M, 11 F, 25 ± 4 years) during controlled breathing on a mouthpiece.
Deposition calculated from aerosol bolus measurements between 50 and 500 mL into
a breath with 50 mL increments. Illustrated data are means and standard errors.
4.2.2.2. Extrathoracic Region
The first line of defense for protecting the lower respiratory tract from inhaled particles is the nose
and mouth. The 2004 PM AQCD (U.S. EPA, 2004) concluded that the ET region, especially the nasal
passages, acts as an efficient filter for small ultrafine particles (<0.01 (mi) and larger particles
(>2 (mi dae), reducing the amount of particles within a wide size range that are available for deposition in
the TB and A regions. Newer studies have become available, but are generally limited to computational
fluid dynamics (CFD) modeling and experimental measurements in casts. As most of these studies do not
substantially improve our understanding of deposition in the ET region they are not reviewed here.
For particles >1 |im dae, deposition efficiency in the oral and nasal passages has been generally
described as a function of an impaction parameter (Stokes number) with the addition of a flow regime
December 2008
4-10
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
parameter (Reynolds number) for the oral passages (Finlay and Martin; Grgic et al., 2004; Kelly et al.,
2005; Schroeter et al., 2006). For an adult male, the CFD simulations of Schroeter et al. (2006) predicted
nasal deposition of 10 |im dae particles was 90%, and 100% for a Ve of 7.5 L/min (rest) and 15 L/min
(light activity), respectively. Thus, relatively few large coarse particles will pass through the nasal
passages into the lungs.
Since the nasal passages are more efficient at removing inhaled particles than the oral passage, an
individual's mode of breathing (i.e., oral vs. nasal) influences the quantity of particles penetrating to the
lung. In limited studies, it has been shown that children tend to have more oral breathing both at rest and
during exercise and also displayed more variability than adults (Becquemin et al., 1999; Bennett et al.,
2008; James et al., 1997). In contrast to adults, there is little data on the uptake of particles for oral or
nasal breathing in children. Theoretical calculations by Xu and Yu (1986) predict enhanced deposition of
particles (greater than 2 |im) in the head region for children when compared to adults. Studies of fine
particle deposition in physical models of the nose, scaled to adult vs. children sizes, predict that
deposition efficiency in the nose is a function of pressure drop across the nose (Phalen et al., 1989).
Consequently, these model analyses suggest that, when properly scaled physiological flows are used in
the calculation of nasal deposition, children, who have higher nasal resistance than adults, should have
higher nasal deposition compared to adults. Surprisingly, the few studies reporting measures of nasal
deposition in children, found lower nasal deposition efficiencies for fine particles (1-3 |im dae) as
compared to adults, despite their higher nasal resistances (Becquemin et al., 1991; Bennett et al., 2008).
These findings of lesser nasal vs oral breathing and less efficient nasal deposition suggest that children's
lower respiratory tract may receive a higher dose of ambient PM compared to adults.
4.2.2.3. Tracheobronchial and Alveolar Region
Inhaled particles passing the ET region enter and may become deposited in the lungs. For any
given particle size, the pattern of particle deposition influences clearance by partitioning deposited
material between lung regions. Deposition in the tracheobronchial airways and alveolar region cannot be
directly measured in vivo. Much of the available deposition data for the TB and A regions have been
obtained from experiments with radioactively labeled, poorly soluble particles (U.S. EPA, 1996) or by use
of aerosol bolus techniques (U.S. EPA, 2004). In general, the ability of these experimental data to define
specific sites of particle deposition is limited to anatomically large regions of the respiratory tract such as
the head, larynx, bronchi, bronchioles, and alveolar region. Mathematical modeling can provide more
refined predictions of deposition sites. Comparisons of the modeling results obtained with two publicly
available models were provided in Figures 4-3 and 4-4. Highly localized sites of deposition within the
bronchi are described in Section 4.2.2.4. Both experimental and modeling techniques are based on many
December 2008
4-11
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
assumptions that may be relatively good for the healthy lung but not for the diseased lung. For discussion
of these issues, the reader is referred to Sections 4.2.4.3 and 4.2.4.4.
4.2.2.4. Localized Deposition Sites
From a toxicological perspective, it is important to realize that not all epithelial cells in an airway
will receive the same dose of deposited particles. Localized deposition in the vicinity of airway
bifurcations has been analyzed using experimental and mathematical modeling techniques. In the 1996
PM AQCD, experimental data were available illustrating the peak deposition of coarse particles (3, 5, and
7 (mi dae) in daughter airways during inspiration and the parent airway during expiration, but always near
the carinal ridge (Kim and Iglesias, 1989; Kim et al., 1989). In the 2004 PM AQCD, mathematical models
predicted distinct "hot spots" of deposition in the vicinity of the carinal ridge for both coarse (10 (j,m) and
ultrafine (0.01 |im) particles (Heistracher and Hofmann, 1997; Hofmann et al., 1996). In a model of
generations 4-5 during inspiration, hot spots occurred at the carinal ridge for 10 |im dae particles due to
inertial impaction and for 0.01 (jm particles due secondary flow patterns formed at the bifurcation. During
expiration, preferential sites of deposition for both particle sizes occurred 1) approaching the juncture of
daughter airways on the walls forming and across the lumen from the carinal ridge and 2) the top and
bottom (visualizing the Y-shaped geometry laying horizontal) of the parent airway downstream of the
bifurcation.
Recent studies further support these findings (Balashazy et al., 2003; Farkas et al., 2006; 2008;
Isaacs et al., 2006). Most of these studies quantified localized deposition in terms of an enhancement
factor. Typically, the enhancement factor is the ratio of the deposition in a pre-specified surface area (e.g.,
100 x 1 ()() jim which corresponds to ~ 10 x 10 epithelial cells) to the average deposition density for the
whole airway geometry. These enhancement factors are very sensitive to the size of the surface
considered (Balashazy et al., 1999). The studies by Farkas et al. (2006; 2008) investigated the phenomena
of localized deposition down to 0.001 (jm particles. The deposition of 0.001 |im was rather uniform,
however, the deposition pattern became increasingly less uniform with increasing particle size. These
studies indicate that, for particles greater than -0.01 |im. some cells located near the carinal ridge of
bronchial bifurcations may receive hundreds to thousands times the average dose (particles per unit
surface area) of the parent and daughter airways. Furthermore, the inertial impaction of particles > 1 |im
dae at the carinal ridge of large bronchi will increase with increasing inspiratory flows. In a comparison of
constricted versus healthy airways, Farkas et al. (2006) also reported that the overall deposition efficiency
of 10 |im dae particles at bifurcations downstream of a constriction may be increased by 18 times. Given
these considerations, Phalen and Oldham (2006) noted that substantial doses of particles (> 1 (im dae) may
be justified for in vitro studies using tracheobronchial epithelial cell cultures.
December 2008
4-12
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
4.2.3. Interspecies Patterns of Deposition
The primary purpose of this document is to assess the health effects of particles in humans. As
such, human dosimetry studies have been stressed in this chapter. Such studies avoid the uncertainties
associated with the extrapolation of dosimetric data from laboratory animals to humans. However, animal
models have been and continue to be used in evaluating PM health effects because of ethical
considerations regarding the types of studies that can be performed with human subjects. Thus, there is a
considerable need to understand dosimetry in animals and dosimetric differences between animals and
humans. Limited new data are becoming available. Similar deposition efficiencies have been reported in
nasal casts of human and rhesus monkey for 1 to 10 (j,m dae for inspiratory flows mimicking resting
breathing patterns (Kelly et al., 2005). Oldham and Robinson (2007) recently provided morphologic data
and predicted particle deposition in an asthma mouse model.
Interspecies similarities and differences in deposition were described in detail in the last two PM
AQCDs (U.S. EPA, 1996, 2004). It was concluded that the general pattern of total particle deposition
efficiency was similar between laboratory animals and humans: deposition increases on both sides of a
minimum that occurs for particles of 0.2 to 1 |im. There are, however, marked interspecies differences in
uptake into the respiratory tract and regional deposition. For instance, the nasal inhalability of 10 ^m dae
particles is predicted to be 96% in humans, whereas it is only 44% in rats (Menache et al., 1995). In most
laboratory animal species (rat, mouse, hamster, guinea pig, and dogs), deposition in the ET region is near
100% percent for particles greater than 5 (jm dae (Raabe et al., 1985), indicating greater efficiency than
that seen in humans. Detailed presentation of dosimetric difference between rats and humans are available
elsewhere (Brown et al., 2005; Jarabek et al., 2005).
Brown et al. (2005) conducted a thorough evaluation of extrapolations between rats and humans in
relation to particulate exposures. One of many factors they considered was the choice of a dose metric
appropriate for comparison between species. For example, deposited mass may be an appropriate PM
indicator for health effects associated with soluble PM constituents. For health effects associated with
insoluble PM, the particle number, surface area, or mass may be appropriate indicators. Given
interspecies differences in deposition patterns and clearance rates, the question of retained versus
deposited dose was also discussed. It was concluded that for acute effects, the maximum deposited
incremental dose may be the appropriate type of dose metric. For chronic effects, long-term burden may
be more appropriate. For various dose metrics, estimates of particle concentration and exposure duration
required for a rat to receive the same dose as received by a human were obtained with consideration of
activity levels and particle size distributions. It was noted that high PM exposures over the period of
months can lead to particle overload in rats (see Section 4.3.4.4). Exposure regimes were derived as a
December 2008
4-13
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
function of particle size and exposure duration that should avoid overwhelming macrophage mediated
clearance achieving particle overload in rats (see Table 12, Brown et al., 2005). Their dosimetric
calculations indicated that to achieve nominally similar acute doses per surface area in rats, relative to
humans undergoing moderate to high exertion, PM exposure concentrations for rats would need to be
somewhat higher than for humans. Since particle clearance from the lungs of rats is faster than humans,
much higher exposure concentrations are required for the rat to simulate retained burdens of humans.
Illustrating the complexity of such analyses, in some cases, rats were found to require lower exposures
than humans to have comparable doses.
4.2.4. Biological Factors Modulating Deposition
Evaluation of factors affecting particle deposition is important to help understand potentially
susceptible subpopulations. Differences in biological response following pollutant exposure may be
caused by dosimetry differences as well as by differences in innate sensitivity. The effects of different
biological factors on deposition were discussed in the 2004 PM AQCD (U.S. EPA, 2004) and are
summarized briefly here.
4.2.4.1. Age
Airway structure and respiratory conditions vary with age, and these variations may alter the
amount and site of particle deposition in the respiratory tract. It was concluded in the 2004 PM AQCD
(U.S. EPA, 2004) that significant differences between adults and children had been predicted by
mathematical models and observed in experimental studies. Studies generally indicated that ET and TB
deposition was greater in children and that children received greater doses of particles per lung surface
area than adults. Deposition studies in the elderly are still quite limited.
Breathing patterns are well recognized to change with increasing age, i.e., tidal volumes increase
and respiratory rates decrease (Tabachnik et al., 1981; Tobin, 1983). Bennett and Zeman (1998) measured
deposition fraction of inhaled, fine particles in children as they breathed the aerosol with their natural,
resting breathing pattern. Among the children, variation in deposition fraction, measured by photometry at
the mouth, was highly dependent on intersubject variation in tidal volume. On the other hand, they found
no difference in deposition fraction for the children vs. adults for these fine particles. This finding and the
modeling predictions (Hofmann et al., 1989) are explained in part by the smaller tidal volume and faster
breathing rate of children relative to adults for natural breathing conditions. Bennett et al. (2008) also
recently reported measures of fine particle deposition fraction for ventilation associated with light
exercise in children and adults and showed that, like with resting breathing, deposition fraction was
December 2008
4-14
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
predicted by breathing pattern and was not different or tended to be less in children compared to adults.
On the other hand, because children breathe at higher minute ventilations relative to their lung volumes,
the rate of deposition of fine particles normalized to lung surface area may be greater in children vs.
adults (Bennett and Zeman, 1998).
Bennett and Zeman (2004) expanded their measures of fine particle deposition during resting
breathing to a larger group of healthy children (6-13 yr; 20 boys, 16 girls) and found again that the
variation in total deposition, was best predicted by tidal volume (r = 0.79, p < 0.001). But both tidal
volume and resting minute ventilation increased with both height and body mass index of the children.
Interestingly, these data suggest that for a given height and age, children with higher body mass index
(BMI) have larger minute ventilations and tidal volumes at rest than those with lower BMI. These
differences in breathing patterns as a function of BMI translated into increased deposition of fine particles
in the heaviest children, the rate of deposition (i.e., particles depositing/time) in the overweight children
was 2.8 times that of the leanest children (p < 0.02). Among all children, the rate of deposition was
significantly correlated with BMI (r = 0.46, p < 0.004). Some of the increased deposition fraction in
heavier children may be due to their elevated tidal volume, which was well correlated with BMI (r = 0.72,
p< 0.001).
In 62 healthy adults with normal lung function aged 18-80, Bennett et al. (1996) showed there was
no effect of age on the whole lung deposition fraction of 2-|im particles under natural breathing
conditions. Across all subjects, the deposition fraction was found to be independent of age, depending on
breathing period (r = 0.58, p < 0.001) and airway resistance (r = 0.46, p < 0.001). In the same adults
breathing with a fixed pattern (360 mL tidal volume, 3.4 sec breathing period), there was a mild decrease
in deposition with increasing age, which could be attributed to increased peripheral airspace dimensions
in the elderly.
4.2.4.2. Gender
Males and females differ in body size, conductive airway size, and ventilatory parameters;
therefore, gender differences in deposition might be expected. In some of the controlled studies, however,
the men and women were constrained to breathe at the same tidal volume and frequency. Since women
are generally smaller than the men, the increased minute ventilation compared to their normal ventilation
could affect deposition patterns. This may help explain why gender related effects on deposition have
been observed in some studies.
Kim and Hu (1998) assessed the regional deposition patterns of 1-, 3-, and 5-|im MMAD particles
in healthy adult males and females using controlled breathing. The total fractional deposition in the lungs
was similar for both genders with the 1 -|im particle size, but was greater in women for the 3- and 5-|im
December 2008
4-15
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
particles. Deposition also appeared to be more localized in the lungs of females compared to those of
males. Kim and Jaques (2000) measured deposition in healthy adults using sizes in the ultrafine mode
(0.04 to 0.1 (mi). Total fractional lung deposition was greater in females than in males for 0.04- and 0.06-
|im particles. The region of peak fractional deposition was shifted closer to the mouth and peak height
was slightly greater for women than for men for all exposure conditions. The total and regional deposition
data from these studies are illustrated in Figure 4-5. These differences were generally attributed to the
smaller size of the upper airways, particularly of the laryngeal structure, in females.
In another study (Bennett et al., 1996), the total respiratory tract deposition of 2-(jm particles was
examined in adult males and females aged 18 to 80 years who breathed with a normal resting pattern.
There was a tendency for a greater deposition fraction in females compared to males. However, since
males had greater minute ventilation, the deposition rate (i.e., deposition per unit time) was greater in
males than in females. More recently, Bennett and Zeman (2004) found no difference in the deposition of
2-(im particles in boys versus girls aged 6-13 yr (n = 36).
4.2.4.3. Anatomical Variability
Anatomical variability, even in the absence of respiratory disease, can affect deposition throughout
the respiratory tract. The ET region is the first exposed to inhaled particles and, therefore, deposition
within this region would reduce the amount of particles available for deposition in the lungs. Variations in
relative deposition within the ET region will, therefore, propagate through the rest of the respiratory tract,
creating differences in calculated doses from individual to individual.
The influence of variations in nasal airway geometry on particle deposition has been investigated.
Cheng et al. (1996) examined nasal airway deposition in healthy adults using particles ranging in size
from 0.004 to 0.15 |im and at two constant inspiratory flow rates, 167 and 333 mL/s. Inter-individual
variability in deposition was correlated with the wide variation of nasal dimensions, in that greater surface
area, smaller cross-sectional area, and increasing complexity of airway shape were all associated with
enhanced deposition. Bennett and Zeman (2005) have also shown that nasal anatomy influences the
efficiency of particle uptake in the noses of adults. For light exercise breathing conditions in adults, their
study demonstrated that nasal deposition efficiencies for both 1- and 2-|im monodisperse particles were
significantly less in African Americans versus Caucasians. The lesser nasal efficiencies in African-
Americans were associated with both lower nasal resistance and less elliptical nostrils compared to
Caucasians.
Within the lungs, the branching structure of the airways may also differ between individuals. Zhao
et al. (2008) recently examined the bronchial anatomy of the left lung in patients (132 M, 84 W; mean age
47 years). At the level of the segmental bronchus in the upper and lower lobes, a bifurcation occurred in
December 2008
4-16
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
the majority of patients. A trifurcation, however, was observed in 23% of the upper and 18% of the lower
lobes. Other more unusual findings were also reported such as four bronchi arising from the left upper
lobe bronchus. As described in Section 4.2.2.4, deposition can be highly localized near the carinal ridge of
bifurcations. The effect of a bifurcation versus other branching patterns on airflow patterns and particle
deposition has not been described in the literature. Martonen et al. (1994) showed that a wide blunt
carinal ridge shape dramatically affected the flow stream lines relative to a narrower and more rounded
ridge shape. Specifically, there were high flow velocities across the entire area of the blunt carinal ridge
versus a smoother division of the airstream in the case of the narrow rounded ridge shape. The implication
may be that localized particle deposition on the carinal ridge would increase with ridge width. A similar
situation might be expected for a trifurcation versus a bifurcation. These differences in branching patterns
provide a clear example of anatomical variability between individuals that might affect both air flow
patterns and sites of particle deposition.
4.2.4.4. Respiratory Tract Disease
The presence of respiratory tract disease can affect airway structure and ventilatory parameters,
thus altering deposition compared to that occurring in healthy individuals. The effect of airway diseases
on deposition has been studied extensively, as described in the 1996 and 2004 PM AQCD (U.S. EPA,
1996, 2004). Studies described therein showed that people with COPD had very heterogeneous deposition
patterns and differences in regional deposition compared to healthy individuals. People with obstructive
pulmonary diseases tended to have greater deposition in the TB region than did healthy people.
Furthermore, there tended to be an inverse relationship between bronchoconstriction and the extent of
deposition in the A region, whereas total respiratory tract deposition generally increased with increasing
degrees of airway obstruction.
The vast majority of deposition studies in individuals with respiratory disease have been performed
during controlled breathing, i.e., all subjects breathed with the same tidal volume and respiratory rate.
However, although resting tidal volume is similar or elevated in people with COPD compared to healthy
individuals, the former tend to breathe at a faster rate, resulting in higher than normal tidal peak flow and
resting minute ventilation. Thus, given that breathing patterns differ between healthy and obstructed
individuals, particle deposition data for controlled breathing may not be appropriate for estimating
respiratory doses from ambient PM exposures.
Bennett et al. (1997) measured the fractional deposition of insoluble 2-|im particles in moderate-to-
severe COPD patients (n = 13; mean age 62 years) and healthy older adults (n = 11; mean age 67 years)
during natural resting breathing. COPD patients had about a 50% greater deposition fraction and a 50%
increase in resting minute ventilation relative to the healthy adults. As a result, the patients had an average
December 2008
4-17
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
deposition rate of about 2.5 times that of healthy adults. Similar to previously reviewed studies
(U.S. EPA, 1996, 2004), these investigators observed an increase in deposition with an increase in airway
resistance, suggesting that deposition increased with the severity of airway disease.
Brown et al. (2002) measured the deposition of an ultrafine aerosol (CMD = 0.033 |im) during
natural resting breathing in 10 patients with moderate-to-severe COPD (mean age 61 years) and 9 healthy
adults (mean age 53 years). The COPD group consisted of 7 patients with chronic bronchitis and 3
patients with emphysema. The aerosol deposition fraction in the bronchitic patients (0.67) was
significantly (p < 0.02) greater than in either the patients with emphysema (0.48) or the healthy subjects
(0.54). Minute ventilation increased with disease severity (healthy, 5.8 L/min; chronic bronchitic, 6.9
L/min; emphysema, 11 L/min). Relative to the healthy subjects, the average dose rate was significantly (p
< 0.05) increased by 1.5 times in the COPD patients, whereas the average deposition fraction only tended
to be increased by 1.1 times. These data further demonstrate the need to consider dose rates (which
depend on minute ventilation) rather than just deposition fractions when evaluating the effect of
respiratory disease on particle deposition and dose.
0.8
Hygroscopic
.2 0.6
t;
LL
0.4
£=
O
-i—¦
if)
O
§" 0.2
Q
Hydrophobic
0.0
0.03
0.1
0.4
Inhaled particle diameter (pm)
Source: Adapted from Tu and Knutson (1984).
Figure 4-6. Total deposition of hygroscopic sodium chloride and hydrophobic aluminosilicate
aerosols during oral breathing (Vt= 1.0 L, f = 15 min1).
4.2.4.5. Hygroscopicity of Aerosols
Experimental and modeling studies of hygroscopic aerosol growth and deposition in the lung were
extensively discussed in Section 10.4.3.1 of the 1996 PM AQCD. Hygroscopic ambient aerosols include
sulfates, nitrates, some organics, and aerosols laden with sodium or potassium. The high relative humidity
in the lungs contributes to rapid growth of hygroscopic particles and dramatically alters the deposition
December 2008
4-18
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
characteristics of ambient hygroscopic aerosols relative to nonhygroscopic aerosols. Nonhygroscopic
particles in the range of 0.3 |im have minimal intrinsic mobility and low total deposition in the lungs.
However, a 0.3 |im salt particle (dry) will grow in vivo to nearly 2 |im and deposit to a far greater extent
(Anselm et al., 1990). The hygroscopic growth of particles in the respiratory tract decreases diffusive
deposition and increases aerodynamic deposition as illustrated in Figure 4-6.
4.2.5. Summary
Particle deposition in the respiratory tract occurs predominantly by diffusion, impaction, and
sedimentation. Deposition is minimal for particle diameters in the range of 0.1 to 1.0 |im. where particles
are small enough to have minimal sedimentation or impaction and sufficiently large so as to have minimal
diffusive deposition. In humans, total respiratory tract deposition approaches 100% for particles of
roughly 0.01 |im (physical diameter) due to diffusive deposition and for particles of around 10 |im dae due
to the efficiency of sedimentation and impaction.
The first line of defense for protecting the lower respiratory tract from inhaled particles is the
nose and mouth. Nasal deposition approaches 100% in humans for 10 |im dae particles. Experimental
studies show lower nasal particle deposition in children than adults. Relative to adults, children also tend
to breathe more through their mouth which is less efficient for removing inhaled particles than the nose.
These findings suggest that the lower respiratory tract of children may receive a higher dose of ambient
PM compared to adults. Since children breathe at higher minute ventilations relative to their lung
volumes, the rate of particle deposition normalized to lung surface area may be further increased relative
to adults.
People with chronic obstructive pulmonary disease (COPD) generally have greater total
deposition and more heterogeneous deposition patterns compared to healthy individuals. The observed an
increase in deposition correlates with increases in airway resistance, suggesting that deposition increases
with the severity of airway disease. COPD patients also have an increased resting minute ventilation
relative to the healthy adults. This demonstrates the need to consider dose rates (which depend on minute
ventilation) rather than just deposition fractions when evaluating the effect of respiratory disease on
particle deposition and dose.
Modeling studies indicate that, for particles greater than -0.01 |im. some cells located near the
carinal ridge of bronchial bifurcations may receive hundreds to thousands times the average dose
(particles per unit surface area) of the parent and daughter airways. The inertial impaction of particles > 1
(jm dae at the carinal ridge of large bronchi increases with increasing inspiratory flows. Airway
constriction can further augment the overall deposition efficiency of coarse particles at downstream
December 2008
4-19
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
bifurcations. These findings suggest that substantial doses of particles (> 1 (jm dae) may be justified for in
vitro studies using tracheobronchial epithelial cell cultures.
Our ability to extrapolate between species has not generally changed since the 2004 PM AQCD.
However, some considerations related to coarse particles warrant comment. The inhalability of particles
having dae of 2.5, 5, and 10 |im is 80, 65, and 44% in rats, respectively, whereas it remains near 100% for
a dae of 10 |im in humans. In most laboratory animal species (rat, mouse, hamster, guinea pig, and dogs),
deposition in the extrathoracic region is near 100% percent for particles greater than 5 |im dae. By
contrast, in humans nasal deposition approaches 100% for 10 ^m dae. Oronasal breathing versus obligate
nasal breathing further contributes to greater penetration of coarse particles into the lower respiratory tract
of humans than rodents.
4.3. Clearance of Poorly Soluble Particles
This section discusses the clearance and translocation of poorly soluble particles that have
deposited in the respiratory tract. The term "clearance" is used here to refer to the processes by which
deposited particles are removed by mucociliary action or phagocytosis from the respiratory tract.
"Translocation" is used here mainly to refer to the movement of free particles across cell membranes and
to extrapulmonary sites. In the literature, translocation may also refer to the extra- and intracellular
dissolution of particles and the subsequent transfer of dissociated material to the blood through extra- and
intracellular fluids and across the various cell membranes and lung tissues. The clearance and distribution
of soluble particles and soluble constituents of particles are discussed in Section 4.4.
A basic overview of biological mechanisms and clearance pathways from various regions of the
respiratory tract are presented in the following sections. Then regional kinetics of particle clearance are
addressed. Subsequently, an update on interspecies patterns and rates of particle clearance is provided.
The translocation of ultrafine particles is also discussed. Finally, information on biological factors that
may modulate clearance is presented.
4.3.1. Clearance Mechanisms and Kinetics
For any given particle size, the deposition pattern of poorly soluble particles influences clearance
by partitioning deposited material between lung regions. Tracheobronchial clearance of poorly soluble
particles in humans, with some exceptions, is thought (in general) to be complete within 24 to 34 h
through the action of the mucociliary escalator. Clearance of poorly soluble particles from the alveolar
region is a much slower process which may continue from months to years.
December 2008
4-20
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
4.3.1.1. Extrathoracic Region
Particles deposited in either the nasal or oral passages are cleared by several mechanisms. Particles
depositing in the mouth may generally be assumed to be swallowed or removed by expectoration.
Particles deposited in the posterior portions of the nasal passages are moved via mucociliary transport
towards the nasopharynx and swallowed. Mucus flow in the most anterior portion of the nasal passages is
forward, toward the vestibular region where removal occurs by sneezing, wiping, or nose blowing.
4.3.1.2. Tracheobronchial Region
Mucociliary clearance in the tracheobronchial region has generally been considered to be a rapid
process that is relatively complete by 24 to 48 h post-inhalation in humans. Mucociliary clearance is
frequently modeled as a series of "escalators" moving material proximally from one generation to the
next. As such, the removal of particles from an airway generation increases with increasing tracheal
mucus velocity. Assuming continuity in the amount of mucus between airway generations, mucus
velocities decrease and transit times within an airway generation increase with distal progression.
Although clearance from the TB region is generally rapid, experimental evidence discussed in the
1996 and 2004 PM AQCD, showed that a fraction of material deposited in the TB region is retained much
longer. The slow cleared fraction from the TB region was thought to increase with decreasing particle
size. For instance, Roth et al. (1993) showed approximately 93% retention of ultrafine particles (30 nm
median diameter) thought to be deposited in the tracheobronchial region at 24 h post inhalation. The slow
phase clearance of these ultrafine particles continued with an estimated half-time (ti/2) of around 40 days.
Using a technique to target inhaled particles (monodisperse 4.2 (.im MMAD) to the conducting airways,
Moller et al. (2004) observed that 49 ± 9% of particles cleared rapidly (ti/2 of 3.0 ± 1.6 h), whereas the
remaining fraction cleared considerably slower of 109 ± 78 days). A portion of the slow cleared
fraction from the TB region appears to be associated with small bronchioles. For large particles
(dae = 6.2 |im) inhaled at very slow rate to theoretical deposit mainly in small ciliated airways, 50% had
cleared by 24-h post-inhalation. Of the remaining particles, 20% of cleared with a ti/2 of 2.0 days and 80%
with ati/2 of 50 days (Falk et al., 1997). Using the same techniques, Svartengren et al. (2005) also
reported the existence of long-term clearance in humans from the small airways. It should be noted that
the clearance rates for the slow cleared TB fraction still exceeds the clearance rate of the alveolar region
in humans. The underlying sites and mechanisms of long-term TB retention in the smaller airways were
not known and remain unknown.
December 2008
4-21
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
4.3.1.3. Alveolar Region
The primary alveolar clearance mechanism is macrophage phagocytosis and migration to terminal
bronchioles where the cells are cleared by the mucociliary escalator. Alveolar macrophages originate from
bone marrow, circulate briefly as monocytes in the blood, and then become pulmonary interstitial
macrophages before migrating to the luminal surfaces. Under normal conditions, a small fraction of
ingested particles may also be cleared through the lymphatic system. This may occur by transepithelial
migration of alveolar macrophage following particle ingestion or free particle translocation with
subsequent uptake by interstitial macrophages. Snipes et al. (1997) have also demonstrated the
importance of neutrophils in clearance of particles from the alveolar region. Rates of alveolar clearance of
poorly soluble particles vary between species and are briefly discussed in Section 4.3.2. The translocation
of particles from their site of deposition is discussed in Section 4.3.3.
On the basis of in vitro studies, the efficiency of macrophage phagocytosis is thought to be greatest
for particles between 1.5 and 3 |im (Oberdorster, 1988). The decreased efficiency of alveolar macrophage
for engulfing ultrafines increases the time available for these particles to be taken up by epithelial cells
and moved into the interstitium (Ferin et al., 1992). Consistent with this supposition (i.e., translocation
increases with time), an increase in Ti02 transport to lymph nodes has been reported following inhalation
of a cytotoxin to macrophages (Greenspan et al., 1988). Interestingly, the long-term clearance kinetics of
the poorly soluble ultrafine (15-20 nm CMD) iridium (Ir) particles were found to be similar to the kinetics
reported in the literature for micrometer-sized particles (Semmler-Behnke et al., 2007; Semmler et al.,
2004). Semmler-Behnke et al. (2007) concluded that ultrafine Ir particles are less phagocytized by
alveolar macrophage than larger particles, but are effectively removed from the airway surface into the
interstitium. Particles are then engulfed by interstitial macrophages which then migrate to the airway
lumen and are removed by mucociliary clearance to the larynx. The major role of macrophage-mediated
clearance was supported by lavage of relatively few free particles versus predominantly phagocytized
particles at time-points of up to 6-months. It is also possible that some free particles as well as particle-
laden macrophage were carried from interstitial sites via the lymph flow to bronchial and bronchiolar
sites, including bronchial-associated lymphatic tissue, where they were excreted again into the airway
lumen.
4.3.2. Interspecies Patterns of Clearance and Retention
There are differences between species in both the rates of particle clearance from the lung and
manner in which particles are retained in the lung. For instance, based of models of mucociliary clearance
from un-diseased airways, >95% of particles deposited in the tracheobronchial airways of rats are
December 2008
4-22
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
predicted to be cleared by 5 h post deposition, whereas it takes nearly 40 h for comparable clearance in
humans (Hofmann and Asgharian, 2003). As noted in Section 4.3.1.2, however, there is considerable
evidence that a sizeable fraction of particle deposited in the ciliated airways of humans (as well as
canines) are cleared at a far slower rate. In contrast, studies of mice and rats show negligible long-term
retention of particles in the ciliated airways (Kreyling et al., 2006).
1.00
Dog
Human
Guinea Pig
+¦»
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Donaldson, 1990). Rats chronically exposed to high concentrations of insoluble particles experience a
reduction in their alveolar clearance rates and an accumulation of interstitial particle burden (Bermudez et
al., 2002; 2004; Ferin et al., 1992; Oberdorster et al., 1994a; 1994b; Warheit et al., 1997). The influence
of exposure concentration on the pattern of particle retention in rats (exposed to diesel soot) and humans
(exposed to coal dust) was examined by (Nikula et al., 2001). In rats, the DE particles were found to be
primarily in the lumens of the alveolar duct and alveoli; whereas in humans, retained dust was found
primarily in the interstitial tissue within the respiratory acini.
4.3.3. Particle Translocation
Mucociliary and macrophage mediated clearance of poorly soluble particles from the respiratory
tract was discussed in Section 4.3.1. and Section 4.3.2. There is evidence that particles may cross cell
membranes and move from their site of deposition by other mechanisms. The following subsections
discuss the movement of particles from the luminal surfaces of the alveolar region and from the olfactory
region to the brain.
4.3.3.1. Alveolar Region
Numerous studies have examined the translocation of ultrafine particles from their site of
deposition in the lung. Traditionally viewed as a relatively inert particle type, ultrafine Ti02 has received
the most study. At the time the 2004 PM AQCD was released, there were conflicting results regarding the
rate and magnitude of ultrafine carbon translocation from the human lung. Since that time, it has become
well-established that the transport of ultrafine carbon particles from the lung is far slower than that of
soluble materials. However, it has also been shown that ultrafine particles cross cell membranes by
mechanisms different from larger (~1 (j,m) particles and that a fraction of these particles enter capillaries
and may distribute systemically. Details of selected new studies investigating the disposition of poorly
soluble particles are provided in Annex B.
There has been some contention regarding ability of ultrafine carbon particles to rapidly diffuse
from the lungs into the systemic circul ation. Based on their study of 5 healthy volunteers, Nemmar et al.
(2002) suggested that ultrafine carbon particles (< 100 nm) pass rapidly into the systemic circulation.
However, Brown et al. (2002) found that the majority of ultrafine carbon particles (CMD, 33 ± 2 nm)
were still in the lungs of healthy human adult volunteers (n = 9; aged 40 to 67 years) and COPD patients
(n = 10; 45 to 70 years) at 24-h post inhalation. Brown et al. (2002) and Burch (2002) contended that the
findings reported by Nemmar et al. (2002) were consistent with soluble pertechnetate clearance, but not
insoluble ultrafine carbon particles. Highly soluble in normal saline, pertechnetate clears rapidly from the
December 2008 4-24 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
lung with a half-time of -10 mins and accumulates most notably in the bladder, stomach, thyroid, and
salivary glands. Three recent studies have confirmed that the majority (>95%) of ultrafine carbon
particles deposited in the lungs of human volunteers are retained at 24 h post inhalation (Mills et al.,
2006; Wiebert et al., 2006a; 2006b). Wiebert et al. (2006b) modified their aerosol generation system to
reduce leaching of the "mTc radiolabel from carbon particles. Except for a small amount of radiotracer
leaching from particles (1.0 ± 0.6% of initially deposited activity in urine by 24 h), these investigators
found negligible radiolabel and associated particle clearance from the lungs by 70 h. The available data
show that there is not a rapid or significant amount of ultrafine carbon particle migration into circulation
(2002; 2002; Mills et al., 2006; Moller et al., 2008; Wiebert et al., 2006a; 2006b).
Although human studies show that the vast majority of ultrafine carbon particles are retained in the
lungs until at least 24 h post inhalation, both in vitro and in vivo studies support the rapid [< 1 h]
translocation of free ultrafine Ti02 particles across pulmonary cell membranes (Churg et al., 1998; Ferin
et al., 1992; Geiser et al., 2005). Peculiar to Ti02 aerosols, there is evidence that particle aggregates may
disassociate once deposited in the lungs. This disassociation makes inhaled aggregate size the determinant
of deposition amount and site, but primary particle size the determinant of subsequent clearance
(Bermudez et al., 2002; Ferin et al., 1992; Takenaka et al., 1986). Following disaggregation, the ultrafine
Ti02 particles are cleared more slowly and cause a greater inflammatory response (PMN influx) than fine
Ti02 particles (Bermudez et al., 2002; Ferin et al., 1992; Oberdorster et al., 1994a; 1994b; 2000). The
differences in inflammatory effects and possibly lymph burdens between fine and ultrafine Ti02 in many
studies appear related to lung burden in terms of particle surface area and not particle mass or number
(Oberdorster et al., 1992; 1996; 2000; Tran et al., 2000). More recently, others have noted that particle
surface area is not an appropriate metric across all particle types (Warheit et al., 2006). Surface
characteristics such as roughness can also affect protein binding and potentially clearance kinetics, with
smoother Ti02 surfaces being more hydrophobic (Sousa et al., 2004).
Geiser et al. (2005) conducted a detailed examination of the disposition of inhaled ultrafine Ti02 in
20 healthy adult rats. They found that distributions of particles among lung tissue compartments appeared
to follow the volume fraction of the tissues and did not significantly differ between 1 and 24 h post-
inhalation. Averaging 1- and 24-h data, 79.3 ± 7.6% of particles were on the luminal side of the airway
surfaces, 4.6 ± 2.6% were in epithelial or endothelial cells, 4.8 ± 4.5% were in connective tissues, and
11.3 ± 3.9% were within capillaries. Particles within cells were not membrane bound. It is not clear why
the fraction of particles identified in compartments such as the capillaries did not differ between 1 and 24
h post-inhalation. These findings were consistent with the smaller study of 5 rats by Kapp et al. (2004)
who reported identifying Ti02 aggregates in a type II pneumocyte, a capillary close to the endothelial
cells, within the surface-lining layer close to the alveolar epithelium immediately following a 1-h
December 2008
4-25
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
exposure. These studies effectively demonstrate that some inhaled ultrafine Ti02 particles once deposited
on the pulmonary surfaces can rapidly [< 1 h] translocate beyond the epithelium and even into the
vasculature.
Extrapulmonary translocation has also been described for poorly soluble ultrafine gold and iridium
particles. In male Wistar-Kyoto rats exposed to ultrafine gold particles (5-8 nm), Takenaka et al. (2006)
reported a low but significant fraction (0.03 to 0.06% of lung concentration) of gold in the blood from 1
to 7 days post inhalation. Semmler et al. (2004) also found small but detectable amounts of poorly soluble
Ir particle (15 and 20 nm CMD) translocation from the lungs of male Wistar-Kyoto rats to secondary
target organs like the liver, spleen, brain, and kidneys. Each of these organs contained about 0.2% of
deposited Ir. The peak levels in these organs were found 7 days post inhalation. The translocated particles
were largely cleared from extrapulmonary organs by 20 days and Ir levels were near background at 60
days post inhalation. Particles may have been distributed systemically via the gastrointestinal tract.
Immediately after the 6 h inhalation, 18 ± 5 of the deposited Ir particles had already cleared into the
gastrointestinal tract. After 3 wk, 31 ± 5% of the deposited particles were retained in the lung. By 2 and 6
months post inhalation, lung retention was 17 ± 3 and 7 ± 1%, respectively. The particles appeared to be
cleared predominantly from the peripheral lung via the mucociliary escalator into the GI tract and were
found in feces.
A few recent studies have characterized differences in the behavior of fine and ultrafine particles in
vitro. Geiser et al. (2005) found that both ultrafine and fine (0.025 (.un gold, 0.078 |_im Ti02, and 0.2 |_im
Ti02) particles cross cellular membranes by non-endocytic (i.e. involving vesicle formation) mechanisms
such as adhesive interactions and diffusion, whereas the phagocytosis of larger 1 |_im Ti02 particles is
ligand-receptor mediated. Edetsberger et al. (2005) found that ultrafine particles (0.020 |im polystyrene)
translocated into cells by first measurement (~1 min after particle application). Intracellular agglomerates
of 88-117 nm were seen by 15-20 mins and of 253-675 nm by 50-60 mins after particle application. These
intracellular aggregates were thought to be result from particle incorporation into endosomes or similar
structures since Genistein or Cytochalasin treatment generally blocked aggregate formation. Interestingly,
particles did not translocate into dead cells, rather they attached to the outside of the cell membrane.
Amine- or carboxyl-modified surfaces (46 nm polystyrene) did not affect translocation across cultures of
human bronchial epithelial cells with about 6% regardless of the surface characteristics (Geys et al.,
2006).
4.3.3.2. Olfactory Region
Numerous studies have demonstrated the translocation of soluble and poorly soluble particles from
the olfactory mucosa via the axons to the olfactory bulb of the brain. The vast majority of these studies
December 2008
4-26
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
were conducted in rodents. However, DeLorenzo (1970) observed the rapid (within 30-60 mins)
movement of 50 nm silver-coated colloidal gold particles instilled on the olfactory mucosa into the
olfactory bulb of squirrel monkeys. The specifics of this and other key studies that have investigated the
translocation of particles to the olfactory bulb are provided in Annex B.
Two recent studies reported the movement of ultrafine particles deposited in the olfactory region of
the nose along the olfactory nerve and into the olfactory bulb of the brain in rats. Oberdorster et al. (2004)
exposed rats to ultrafine carbon particles (36 nm CMD, 1.7 og) containing 13C in a whole-body chamber
for 6 h. The distribution of 13C was followed for seven days postexposure. There was a significant
increase in 13C in the olfactory bulb on Day 1 with persistent and continued increased through Day 7.
Elder et al. (2006) exposed rats to Mn oxide (-30 nm equivalent sphere with 3-8 nm primary particles)
via body inhalation exposure for 12 d (6 h/d, 5 d/wk) with both nares open or Mn oxide for 2 d (6 h/d)
with right nostril blocked. After the 12 d exposure via both nostrils, Mn in the olfactory bulb increased
3.5-fold, whereas in the lung Mn concentrations only doubled. After the 2 d exposure with the right
nostril blocked, Mn was found mainly in the left olfactory bulb (2.4-fold increase). These studies suggest
the neuronal uptake and translocation of ultrafine particles without particle dissolution and in the absence
of mucosal injury.
Elder et al. (2006) also addressed the issue of whether solubilization of particles was requisite for
translocation along the olfactory nerve and into the brain. Similar amounts of soluble MnCl2 and poorly
soluble Mn oxide were instilled onto the left naris of anesthetized rats. At 24 h post instillation, similar
amounts of Mn were found in the left olfactory bulb of rats instilled with MnCl2 (8.2 ± 3.6% of instilled)
and Mn oxide (8.2 ± 0.7% of instilled). If solubilization were required for translocation, then a lower
amount of Mn oxide than MnCl2 should have reached the olfactory bulb. Following 14 consecutive days
of aerosol exposure, Dorman et al. (2001) demonstrated that more soluble Mn sulfate reaches the
olfactory bulb and striatum of rat brains than the poorly soluble form of Mn tetroxide. Nonetheless, the
Mn levels were statistically increased in both the olfactory bulb and striatum following exposure to Mn
tetroxide relative to filter air. In a subsequent 13-wk exposure study, Dorman et al. (2004) also
demonstrated that more soluble MnS04 reached the olfactory bulb than was observed for the less soluble
Mn form (hureaulite). Both the soluble and less soluble forms of Mn resulted in statistically increased
levels of Mn in the olfactory bulb relative to air exposed controls. The soluble MnS04 was also observed
to reach the striatum and cerebellum. In addition, Yu et al. (2003) demonstrated increased Mn levels in the
brains of rats exposed to welding-fumes for 60 days, however, the role of transport via the blood is less
clear in this study.
The translocation of zinc and titanium dioxide to the olfactory bulb has also been reported in the
literature. Persson et al. (2003) observed the translocation of Zn to the olfactory bulbs following
December 2008
4-27
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
instillation in both rats and freshwater pike. Wang et al. (2007) reported the translocation of both fine (155
nm) and ultrafine (21 and 71 nm) Ti02 particles. These particles are readily characterized in the literature
as poorly soluble. Interestingly, a qualitative analysis of the data showed that more of the fine Ti02 than
ultrafine Ti02 reached the olfactory bulb. Wang et al. (2007) suggested that a strong hydrophilic character
and propensity for aggregation reduced the translocation of the ultrafine Ti02.
The importance of particle translocation to the brain is not yet understood. Translocation via the
axon to the olfactory bulb has been observed for numerous compounds of varying composition, particle
size, and solubility. Although the rate of translocation is rapid, perhaps less than an hour, the magnitude of
transport remains poorly characterized. With regard to the magnitude of transport, Elder et al. (2006)
found that as much as 8% of both soluble and insoluble forms of Mn were translocated to the olfactory
bulb in rats following intranasal instillation. It is also still unclear to what extent translocation to the
olfactory bulb and other brain regions may vary between species. The olfactory mucosa covers
approximately 50% of the nasal epithelium in rodents versus only about 5% in primates (Aschner et al.,
2005). Additionally, a greater portion of inhaled air passes through the olfactory region of rats relative to
primates (Kimbell, 2006). These differences may predispose rats, more so than humans, to deposition of
particles in the olfactory region with subsequent particle translocation to the olfactory bulb.
4.3.4. Factors Modulating Clearance
4.3.4.1. Age
It was previously concluded that there appeared to be no clear evidence for any age-related
differences in clearance from the lung or total respiratory tract, either from child to adult, or young adult
to elderly (U.S. EPA, 1996, 2006d). Studies showed either no change or some slowing in mucus clearance
with age after maturity. Although some differences in alveolar macrophage function were reported
between mature and senescent mice, no age-related decline in macrophage function had been observed in
humans. A comprehensive review of the recent and older literature supports a decrease in mucociliary
clearance with increasing age beyond adulthood in humans and animals. Limited animal data also suggest
macrophage-mediated alveolar clearance may also decrease with age.
Studies addressing the effects of age on respiratory tract clearance are provided in Annex B. Ho et
al. (2001) demonstrated that nasal mucociliary clearance rates were about 40% lower in old versus young
men and women. Tracheal mucus velocities in elderly (or aged) humans and beagle dogs are about 50%
that of young adults (Goodman et al., 1978; Whaley et al., 1987). Several human studies have
demonstrated decreasing rates of mucociliary particle clearance from the large and small bronchial
December 2008
4-28
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
airways with increasing age (Puchelle et al., 1979; Svartengren et al., 2005; Vastag et al., 1985). Linear
fits to the data show that rapid clearance (within 1 h) from large bronchi and prolonged clearance
(between 1-21 days) from the small bronchioles in an 80 year-old is only about 50% of that in 20 year-old
(Svartengren et al., 2005; Vastag et al., 1985). One study reported that alveolar particle clearance rates
decreased by nearly 40% in old versus young rats (Muhle et al., 1990). Another study has reported that
older rats have an increased susceptibility to pulmonary infection due to altered alveolar macrophage
function and slowed bacterial clearance (Antonini et al., 2001). Although data are somewhat limited, they
consistently show a depression of clearance throughout the respiratory tract with increasing age from
young adulthood in humans and laboratory animals.
4.3.4.2. Gender
Gender was not found to affect clearance rates in prior reviews (U.S. EPA, 1996, 2004). Studies not
included in those reviews also show that human males and females have similar nasal mucus clearance
rates (Ho et al., 2001), tracheal mucus velocities (Yeates et al., 1981), and large bronchial airway
clearance rates (Vastag et al., 1985).
4.3.4.3. Respiratory Tract Disease
At the time of the last two reviews (U.S. EPA, 1996, 2004), it was well recognized that obstructive
airways disease may influence both the site of initial deposition and the rate of mucociliary clearance
from the airways. When deposition patterns are matched, mucociliary clearance rates are reduced in
patients with COPD relative to healthy controls.
Using a bolus technique to target specific lung regions, Moller et al. (2008) examined particle
clearance from the ciliated airways and alveolar region of healthy subjects, smokers, and patients with
COPD. Airway retention after 1.5 h was significantly lower in healthy subjects (89 ± 6%) than smokers
(97 ± 3%) or COPD patients (96 ± 6%). At 24 and 48 h, retention remained significantly higher in COPD
patients (86 ± 6% and 82 ± 6%, respectively) than healthy subjects (75 ± 10% and 70 ± 9%, respectively).
However, these findings are confounded by the more central pattern of deposition in the healthy subjects
than in the smokers and COPD patients. Alveolar retention of particles was similar between the groups at
48 h post-inhalation.
Chen et al. (2006b) investigated the effect of endotoxin on the disposition of particles. Healthy rats
and those pretreated with endotoxin (12 h before particle instillation) were instilled with ultrafine (56.4
nm) or fine (202 nm) particles. In healthy rats, there were no marked differences in lung retention or
systemic distribution between the ultrafine and fine particles. In healthy animals, ultrafine particles were
December 2008
4-29
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
primarily retained in lungs (72 ± 10% at 0.5-2 h; 65 ± 1% at 1 d; 62 ± 5% at 5 d). Initially, there was rapid
particle movement into the blood (2 ± 1% at 0.5-2 h; 0.1 ± 0.1% at 5 d) and liver (3 ± 2% at 0.5-2 h;
1 ± 0.1% at 5 d). At 1 d post-instillation, about 13% of the particles were in the urine or feces. In rats
pretreated with endotoxin, by 2 h post-instillation, the ultrafine particles accessed the blood (5 vs. 2%)
and liver (11 vs. 4%) to a significantly greater extent than fine particles. The endotoxin treated rats also
had significantly greater amounts of ultrafine particles in the blood (5 vs. 2%) and liver (11 vs. 3%)
relative to the healthy rats. This study demonstrates that acute lung injury caused by endotoxin increases
the migration of ultrafine particles into systemic circulation.
4.3.4.4. Particle Overload
Unlike other laboratory animals, rats appear susceptible to "particle overload" effects due to
impaired macrophage-mediated alveolar clearance. Numerous reviews have discussed this phenomenon
and the difficulties it poses for the extrapolation of chronic effects in rats to humans (ILSI, 2000; Miller,
2000; Morrow, 1994; Oberdorster, 1995, 2002). Large mammals have slow pulmonary particle clearance
and retain particles in interstitial tissues under normal conditions, whereas rats have rapid pulmonary
clearance and retain particles in alveolar macrophages (Snipes, 1996). With chronic high doses of
particulate there is a shift in the pattern of dust accumulation and response from that observed at lower
particulate doses in rat lungs (Snipes, 1996; Vincent and Donaldson, 1990). Rats chronically exposed to
high concentrations of insoluble particles experience a reduction in their alveolar clearance rates and an
accumulation of interstitial particle burden (Bermudez et al., 2002; 2004; Ferin et al., 1992; Oberdorster
et al., 1994a; 1994b; Warheit et al., 1997). With continued exposure, some rats eventually develop
pulmonary fibrosis and both benign and malignant tumors (Lee et al., 1985a, b; 1986; Warheit et al.,
1997). Oberdorster (1996, 2002) proposed that high-dose effects observed in rats may be associated with
two thresholds. The first threshold is the pulmonary dose that results in a reduction in macrophage-
mediated clearance. The second threshold, occurring at a higher dose than the first, is the dose at which
antioxidant defenses are overwhelmed and pulmonary tumors develop. Intrapulmonary tumors following
Ti02 exposures are exclusive to rats and are not found in mice or hamsters (Mauderly, 1997). Moreover,
Lee et al. (1985b) noted that the squamous cell carcinomas observed with prolonged high concentration
Ti02 exposures developed from the alveolar lining cells adjacent to the alveolar ducts, whereas squamous
cell carcinomas in humans are generally linked with cigarette smoking are thought to arise from basal
cells of the bronchial epithelium. Quoting Lee et al. (1986), "Since the lung tumors were a unique type of
experimentally induced tumor under exaggerated exposure conditions and have not usually been seen in
man or animals, their relevance to man in questionable."
December 2008
4-30
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
4.3.5. Summary
For any given particle size, the pattern of poorly soluble particle deposition influences clearance by
partitioning deposited material between regions of the respiratory tract. Particles depositing in the mouth
may generally be assumed to be swallowed or removed by expectoration. Particles deposited in the
posterior portions of the nasal passages or the tracheobronchial airways are moved via mucociliary
transport towards the nasopharynx and swallowed. Although clearance from the tracheobronchial region
is generally rapid, there appears to be fraction of material deposited in the tracheobronchial region of
humans that is retained much longer. The underlying sites and mechanisms of long-term tracheobronchial
retention are not known. In contrast to humans, mice and rats appear to have negligible long-term
retention of particles in tracheobronchial airways. The primary alveolar clearance mechanism is
macrophage phagocytosis and migration to terminal bronchioles where the cells are cleared by the
mucociliary escalator. Clearance from both the tracheobronchial and alveolar region is more rapid in
rodents than humans. Mucociliary and macrophage-mediated clearance decreases with age beyond
adulthood.
Human data show that there is not a rapid or significant amount of ultrafine carbon particle
migration into circulation. However, both in vitro and in vivo studies support the rapid [< 1 hr]
translocation of free ultrafine Ti02 particles across pulmonary cell membranes. Extrapulmonary
translocation has also been described in rats for poorly soluble ultrafine gold and iridium particles. A low,
but statistically significant, fraction (0.03 to 0.06% of lung concentration) of ultrafine gold particles has
been observed in the blood of rats from 1 to 7 days post inhalation. The translocation in detectable
amounts (< 1% of deposited material) of poorly soluble Ir particles (15 and 20 nm CMD) from the lungs
of rats to secondary target organs like the liver, spleen, brain, and kidneys has also been reported.
However, the systemic distribution of particles may have occurred via normal clearance from the lungs to
the gastrointestinal tract.
Although the importance of particle translocation to the brain is not yet understood, translocation
from the olfactory mucosa via the axon to the olfactory bulb has been reported in primates, rodents, and
freshwater pike for numerous compounds of varying composition, particle size, and solubility. The rate of
translocation is rapid, perhaps less than an hour. In rats, as much as 8% of material may become
translocated to the olfactory bulb following intranasal instillation. It is unclear to what extent
translocation to the olfactory bulb and other brain regions may vary between species. Interspecies
differences may predispose rats, more so than humans, to the deposition of particles in the olfactory
region with subsequent translocation to the olfactory bulb.
December 2008
4-31
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
4.4. Clearance of Soluble Materials
Soluble particles and soluble constituents of particles may be absorbed through the epithelium and
distributed systemically or retained in the lung. The rate of dissolution depends on a number of factors,
including particle surface area and chemical structure. Some dissolved materials bind to proteins or other
components in the airway surface liquid layer. In the ciliated airways, solutes are cleared by mucociliary
transport and diffuse into underlining tissues and the blood. In the alveolar regions, the thin barrier
between the air and blood allows for rapid transport of solutes into the blood. The movement of soluble
materials depends on the site of deposition in the lung, the rate of material dissolution from particles, and
the molecular weight of the solute. The rate of soluble material clearance from the lungs depends on
epithelial permeability which may be affected by age, respiratory disease, and concurrent exposures.
While enhanced clearance of insoluble particles acts to reduce dose to airway tissue, increased transport
of soluble matter into the blood stream may enhance effects on extra-pulmonary organs.
4.4.1. Clearance Mechanisms and Kinetics
The rate of absorption across the epithelium for materials that dissolve in the airway or alveolar
lining fluid is fairly rapid (minutes to hours) and is a function of their molecular size and their water or
lipid solubility (Enna and Schanker, 1972; Huchon et al., 1987; Oberdorster, 1988; Schanker et al., 1986).
Huchon et al. (1987) studied the clearance of a variety of aerosolized solutes from the lungs of dogs.
Solute clearance was inversely related to molecular weight. Negligible clearance of the largest molecular
weight solute (transferrin; mol wt -76,000 daltons) in their study was found over a 30-min observation
period. At the other extreme, free pertechnetate (mol wt -163 daltons) had a clearance rate of 6% per min.
Clearance ofhydrophilic solutes is diffusion limited by pore sizes associated with intercellular tight
junctions (estimated at 0.6 to 1.5 nm). Absorption of lipophilic compounds that pass easily through cell
membranes is perfusion limited and thus generally occurs very rapidly. However, if lipophilic materials
are adsorbed onto insoluble particles their retention in the lung may be prolonged (Creasia et al., 1976). In
addition to diffusion through intercellular junctions, transcellular transport of large solutes by pinocytosis
in epithelial cells has also been observed (Chinard, 1980). Once a particle is phagocytosed by an alveolar
macrophage it may slowly dissolve and be released from the cell to move across the epithelium into the
bloodstream. The dissolution rate is inversely related to particle size and directly related to specific
surface area (Kreyling and Scheuch, 2000) and facilitated by the acidic (pH of 4.3 to 5.3) environment of
the phagolysosome (Kreyling, 1992).
December 2008
4-32
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
There is considerable evidence as well that soluble particles depositing in the bronchial airways are
also cleared by mucociliary transport (Bennett and Ilowite, 1989; Lay et al., 2003; Matsui et al., 1998;
Sakagami et al., 2002; Wagner and Foster, 2001). The relative contribution of their removal by
transepithelial absorption vs. mucociliary clearance is likely a function of both the molecular size and
water or lipid solubility of the material (Enna and Schanker, 1972; Huchon et al., 1987; Oberdorster,
1988; Sakagami et al., 2002). Furthermore, the rate of mucociliary transport for soluble particles may be
less than that of insoluble particles (Lay et al., 2003). Consequently, non-permeating hydrophilic solutes
may remain in contact with the airway epithelium for a longer period than insoluble particles. This may
be due to diffusion of a greater portion of the solute into the periciliary sol layer which may be
transported less efficiently than the mucus layer during mucociliary clearance. Bronchial blood flow has
also been shown to modulate airway retention of soluble particles (Wagner and Foster, 2001), i.e.
decreasing blood flow increases airway retention of soluble particles.
As an example of how transport of soluble components of PM may clear the lung by transepithelial
absorption, Wallenborn et al. (2007) measured elemental content of lungs, plasma, heart, and liver of
healthy male WKY rats (12-15 weeks old) 4 or 24 h following a single intratracheal (IT) instillation of
saline or 8.33 mg/kg of oil combustion PM containing a variety of transition metals with differing water
and acid solubility. Metals with high water solubility and relatively high concentration in oil combustion
PM were increased in extrapulmonary organs. Elements with low water or acid solubility, like silicon and
aluminum, were not detected in extrapulmonary tissues despite decreased levels in the lung suggesting
they cleared the lung primarily by mucociliary clearance. Thus, PM-associated metals deposited in the
lung may be released into systemic circulation at different rates depending on their water/acid solubility,
thereby providing a means by which metals may elicit direct extrapulmonary effects.
4.4.2. Factors Modulating Clearance
A number of studies have evaluated the epithelial permeability by measuring the clearance of
"mTc-diethylenetriaminepentaacetic acid (99mTc-DTPA), a small hydrophilic solute (492 daltons, 0.57
nm). These studies are the basis for much of the discussion in this section.
4.4.2.1. Age
In humans, the clearance of water-soluble particles (99mTc-DTPA) from the alveolar epithelium
generally slows with increasing age (Braga et al., 1996; Pigorini et al., 1988). However, Tankersley et al.
(2003) recently showed enhanced permeability of soluble particles (99mTc-DTPA) in terminally senescent
December 2008
4-33
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
mice just before death, suggesting that a disintegration of the epithelial barrier may be a feature of lung
homeostatic loss during this period of terminal senescence.
4.4.2.2. Exercise
The transepithelial transport rates of soluble particles, "mTc-DTPA, have also been found to
increase during exercise (Hanel et al., 2003; Lorino et al., 1989; Meignan et al., 1986). This enhancement
was linked to increases in tidal volume associated with exercise (Lorino et al., 1989). Regionally, this
effect was dominated by increased apical lung clearance and attributed to an increase in apical blood flow
(Meignan et al., 1986). The increased permeability with exercise appears to resolve to baseline after a
short period post exercise, i.e. within a couple hours (Hanel et al., 2003).
4.4.2.3. Disease
Because the integrity of the epithelial surface lining of the lungs may be damaged from lung
disease, particles (either insoluble or soluble) may gain greater access to the interstitium, lymph, and
blood stream. Damage to the epithelial barrier is most likely to acutely affect transepithelial transport
rates of soluble particles. From bronchial biopsies, Laitinen et al. (1985) found various degrees of
epithelial damage, from loosening of tight junctions to complete denudation of the airway epithelium, in
asthmatics. Consistent with these findings, Ilowite et al. (1989) found that asthmatics had increased
permeability of the bronchial mucosa to the hydrophilic solute "mTc-DTPA. On the other hand, a more
recent study in a sheep model showed that the presence of bronchial edema could slow the uptake of
soluble DTPA into the blood and enhanced retention in the airways, likely within the expanded interstitial
barrier (Foster and Wagner, 2001). Both a leaky epithelial barrier and expanded interstitial barrier
associated with asthma may result in enhanced exposure of submucosal immune and smooth muscle cells
to xenobiotic substances.
Alveolar epithelial permeability was also shown affected by the presence of lung inflammation.
The most common finding has been a clear increase in alveolar permeability induced by cigarette
smoking (Jones et al., 1980). This effect appears to be dependent the recent cigarette smoke exposure as
indexed by carboxyhaemoglobin (Jones et al., 1983) and is rapidly reversible within a week smoking
cessation (Mason et al., 1983). In fact, Huchon et al. (1984) demonstrated that COPD patients who have
stopped smoking have normal clearance of "mTc-DTPA.
In general, increased alveolar permeability to "mTc-DTPA has been found to be associated with any
lung syndrome characterized by pulmonary edema. While the trans-alveolar transport of a small solute
like DTPA is very sensitive to even mild acute lung injury (such as that associated with even mild
December 2008
4-34
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
cigarette smoking), increased transport rates of larger molecules (>100K daltons) across the alveolar
epithelium require more severe damage like that seen in adult respiratory distress syndrome (ARDS)
(Braude et al., 1986; Peterson et al., 1989). Interstitial lung disease and pulmonary fibrosis are also
characterized by increased alveolar permeability (Antoniou et al., 2006; Bodolay et al., 2005; Watanabe et
al., 2007). Interestingly, these recent studies have also shown that the increased permeability in these
patients could be corrected with immunosuppressive/steroid treatments (Bodolay et al., 2005; Watanabe et
al., 2007). Furthermore, studies of DTPA clearance in bleomycin injured dogs, a model of pulmonary
fibrosis, suggest that the enhanced permeability is associated with the initial acute phase of the lung
damage, with clearance rates returning to normal as chronic fibrosis developed over time (Suga et al.,
2003).
Finally, as evidence of lung complications associated with non-insulin dependent diabetes (type 2)
patients, Lin et al. (2002a) found impairment of alveolar integrity as shown by increased transport rates of
both hydophilic and lipophilic solutes from the lungs in these patients. By contrast, a number of other
studies have found epithelial permeability reduced, i.e. slower transport rates, in diabetes (Caner, 1994;
Mousa et al., 2000; Ozsahin et al., 2006) that may be related to disease duration and metabolic control
(Ozsahin et al., 2006). These findings are consistent with thickening of alveolar basement membrane
detected in autopsies of diabetes patients (Weynand et al., 1999).
4.4.2.4. Concurrent Exposures
The integrity of the alveolar epithelium may be disrupted by copollutants such that soluble
components of inhaled particles can more easily enter the interstitium and blood stream. Like active
cigarette smoking discussed previously, Beadsmoore et al. (2007) showed clearance half-times in healthy
passive smokers to be shorter compared with healthy non-smokers but still longer than in healthy
smokers. These findings show a progressive increase in epithelial permeability with exposure to cigarette
smoke. Similarly, acute exposure of humans to 0.4 ppm ozone for 2 h with intermittent exercise has been
shown to alter epithelial integrity and increase clearance of soluble hydrophilic particles from the alveolar
surfaces of the lung (Kehrl et al., 1987). This effect persists to at least 24 h post-exposure to even low
concentrations (0.24 ppm average for 130 minutes) of ozone (Foster and Stetkiewicz, 1996). Similarly,
0.8 ppm 03 exposure for 2 h in rats shows increased permeability to macromolecules at all levels of the
respiratory tract (Bhalla et al., 1986) that persisted in the alveolar region beyond 24-h post-exposure.
Chang et al. (2005a) recently showed that ultrafine carbon black acts through a reactive oxygen species
(ROS) dependent pathway to increase epithelial permeability in mice.
But chronic exposure to other particulate or gaseous pollutants has not always led to increased
epithelial permeability. Studying subjects with a variety of occupational exposures, Kaya et al. (2006)
December 2008
4-35
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
showed that nonsmoking welders actually have decreased epithelial permeability relative to nonsmoking
control subjects, and occupational exposure of painters to isocyanates has no effect on bronchoalveolar
epithelial permeability (Kaya et al., 2003).
4.4.3. Summary
The healthy airway and alveolar epithelium is generally impermeable to very large insoluble
macromolecules and particles. Water and acid soluble particles may more rapidly move through the
epithelium as they dissolve on the airway surface or within the phagolysomes of macrophages. The
presence of airway inflammation in a variety of airway diseases (e.g. asthma, fibrosis, ARDS, pulmonary
edema, inflammation from smoking) alters epithelial integrity to allow more rapid movement of these
solutes into the bloodstream. While diabetics are another group recently shown to have increased
susceptibility to particulate air pollution (Zanobetti and Schwartz, 2002), it is unclear whether transport of
soluble particles across the epithelium is affected in these patients. In general, it appears that co-exposure
to irritant pollutants results in a disruption of epithelial integrity and macrophage function which, on the
one hand, retards mucociliary and alveolar clearance, but also allows for a more rapid movement of
soluble constituents across the epithelial surface into the interstitium and blood stream. Cohen et al.,
el 997 #9299} may have best illustrated the competing effects of mucociliary and transepithelial transport
by showing that co-exposure to ozone affected the retention of inhaled chromium in rats differently
depending on its solubility. In its soluble potassium chromate form, ozone decreased the retention of
chromium, but when chromium was inhaled as insoluble barium chromate, its retention in the lung was
increased by ozone co-exposure. Similarly, a study that showed decreased clearance of insoluble cesium
oxide particles following influenza infection also showed a virus induced enhancement of clearance for a
soluble cesium chloride (Lundgren et al., 1978). These alterations in epithelial permeability by disease,
pollutant exposure, or infection may partially explain increased susceptibility to PM associated with these
co-conditions.
December 2008
4-36
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Chapter 5. Possible Pathways/
Modes of Action
The mechanisms underlying pulmonary effects of inhaled PM have been well-studied and there is
general agreement regarding the key roles played by cellular injury and inflammation. These pathways
are initiated following the inhalation of particles and their deposition on respiratory tract surfaces. Since
most of these studies were conducted at concentrations of PM higher than ambient levels, there is some
question regarding the relevance of these responses and mechanisms to ambient exposures.
Interestingly, inhaled PM may also affect the cardiovascular, hematopoietic and other systems.
Mechanisms underlying these extra-pulmonary effects are incompletely understood. However, pulmonary
inflammation can lead to systemic inflammation and pulmonary reflexes can activate the autonomic
nervous system. These latter responses may mediate cardiovascular and other systemic effects, as will be
discussed below. In addition, it has been proposed that PM or soluble components of PM reach the
circulation by translocating across the epithelial barrier of the respiratory tract. In this way PM or its
components may interact directly with cells in the vasculature, blood, heart and other organs. At this time,
evidence clearly supports the translocation of soluble components following some high dose exposures
involving intratracheal instillation; however there is insufficient evidence to support translocation of
soluble components or intact particles following inhalation exposures at lower concentrations (see
Chapter 4). Future studies will be required to resolve these issues.
The following sections discuss biological pathways which comprise proposed modes of action for
the pulmonary and extra-pulmonary effects of inhaled PM. Overall themes are emphasized and supportive
evidence from new in vitro and in vivo animal studies is cited. The characterization of evidence here is
for PM in general, since most of the potential pathways or modes of action described below do not appear
to be specific to a particular size class of PM. Finally, a compilation of results from new inhalation studies
which are relevant to ambient PM exposures and which confirm and extend these proposed mechanisms
is found at the end of this chapter. Detailed descriptions of these key new studies are found in Chapters 6
and 7.
December 2008
5-1
DRAFT-DO NOT QUOTE OR CITE
-------
5.1. Pulmonary Effects
5.1.1. Reactive Oxygen Species
1 A great deal of research interest has focused on the role of ROS in the initiation of pulmonary
2 injury and inflammation following exposure to PM. Numerous studies have demonstrated PM oxidative
3 potential in in vitro assay systems (Ayres et al., 2008; Cho et al., 2005a; Shi et al., 2003; Tao et al., 2003).
4 Both redox active surface components such as metals and organic species and the surface characteristics
5 of crystal structures have been shown to contribute to oxidative potential (Jiang et al., 2008; Tao et al.,
6 2003; Warheit et al., 2007). In this way, PM may be a direct source of ROS in the respiratory tract (Figure
7 5-1).
PM Oxidative
Potential
' Cell-free assay or
oxidation of components
v in a cellular system 7
Surface Characteristics
of Crystal Structures
Redox Active
Surface Components
Metals, Organics
Figure 5-1. PM oxidative potential.
8 PM may also act as an indirect source of ROS in the respiratory tract by stimulating cells to
9 produce ROS (Ayres et al., 2008; Tao et al., 2003) (Figure 5-2). This may explain the observation that PM
10 oxidative potential does not always correlate with cellular or tissue oxidative stress. Numerous studies
December 2008
5-2
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
have demonstrated that exposure to PM increases intracellular production of ROS by a variety of
mechanisms. For example, PM interaction with cell surfaces results in stimulation of NADPH oxidase in
macrophages (i.e. the respiratory burst) (Dostert et aL 2008) and in epithelial cells (Amara et al., 2007;
Becher et al., 2007; Tamaoki et aL 2004). Absorption of PM soluble components by respiratory tract cells
can occur (Penn et al,, 2005) and be followed by microsomal transformation of poly cyclic aromatic
hydrocarbons to quinones or by redox cycling of soluble metals with the resulting production of
intracellular ROS (Molinelli et al,, 2002; Xia et al., 2004). Disruption of intracellular iron homeostasis
with the subsequent generation of ROS has also been demonstrated following PM exposure (Ghio and
Cohen, 2005). In some cases, mitochondria serve as the source of ROS in response to PM (Huang et al.,
2003b; Risom et al., 2005; Soberanes et al., 2006). Furthermore, PM interaction with cells can lead to the
induction of nitric oxide synthase (Becher et al., 2007; Lindbom et al., 2007; Xiao et al., 2005; Zhao et
al., 2006a) and the production of nitric oxide and other reactive nitrogen species (RNS).
> f
Cellular Sources
of ROS/RNS
i i
ROS/RNS Assay
Oxidation of Cellular Components
Lipid Peroxidation, Nitrotyrosine
HO-1 Induction
NADPH
Oxidase
Mitochondria)
Electron Transport
Soluble metals
Nitric Oxide
Synthase
Microsomal
Metabolism
(PAH/Quinones)
Binding to Cell Surfaces
Phagocytosis
Cytoskeletal Interactions
Iron Sequestration
Redox Cycling
Altered iron homeostasis
Figure 5-2. PM stimulates pulmonary cells to produce ROS/RNS.
High levels of intracellular ROS/RNS can lead to irreversible protein modifications, loss of cellular
membrane integrity and cellular toxicity. Lower levels of ROS/RNS may involve reversible protein
modifications which trigger intracellular signaling pathways and/or adaptive responses.
December 2008
5-3
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
5.1.2. Activation of Cell Signaling Pathways
Activation of cell signaling pathways by ROS/RNS has received increasing attention in recent
years. An early example was provided by Kaul and Forman (1996) who demonstrated that respiratory
burst-derived H202 activates the transcription factor NFKB. Numerous studies since then have
demonstrated that PM, which serves as both a direct and indirect source of ROS/RNS, activates cell
signaling pathways by this mechanism.
PM exposure results in activation of cell signaling by other mechanisms as well. For example, PM
delivers water-soluble components such as endotoxin and zinc to cell surfaces. Endotoxin binds to toll-
like receptors on alveolar macrophages and results in the upregulation of cytokines (Becker et al., 2002).
Zinc, a transition metal which does not redox cycle, inhibits protein tyrosine phophatases in airway
epithelial cells resulting in a cascade of cell signaling events (Tal et al., 2006). Similarly, PM-mediated
delivery of lipid soluble components such as PAH in mice results in binding and activation of the
arylhydrocarbon receptor (AhR) in mice. AhR is a transcription factor responsible for the upregulation of
CYP1A1, a cytochrome oxidase involved in PAH metabolism (Rouse et al., 2008). In addition, interaction
of PM with cell surfaces by perturbation of the cytoskeleton, adherence, internalization, or
receptor-mediated pathways has been demonstrated to activate cell signaling pathways.
Cell Surface
Interactions
Endotoxin
Zinc
PAH
initiation
Cell Signaling/
Transcription
Facto; Activation
Cytokines
Chemokines
Pro1eases
Mediators
ROS/RNS
Amplification
Amplification
Influx of Leukocytes
Pulmonary Inflammation
Figure 5-3. PM activates cell signaling pathways leading to pulmonary inflammation.
December 2008
5-4
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Recent studies involving PM exposures have focused on intracellular pathways involving protein
kinases such as: MAPK (Aam and Fonnum, 2007; Bayram et al., 2006; Lee et al., 2005a; Roberts et al.,
2003), AKT (Ahsan et al., 2005), src (Cao et al., 2007) and epidermal growth factor receptor (Blanchet et
al., 2004; Cao et al., 2007; Tamaoki et al., 2004); and ras (Tamaoki et al., 2004), toll-like receptors
(Becker et al., 2005a; 2005b), protein tyrosine phosphatases (Tal et al., 2006), phospholipases A2 (Lee et
al., 2003c), calcium (Agopyan et al., 2003; Brown et al., 2004a; 2004b; Geng et al., 2005, 2006;
Sakamoto et al., 2007), caspases (Soberanes et al., 2006; Zhang et al., 2007a), PARP-1 (Zhang et al.,
2007a) and histone acetylation (Gilmour et al., 2003). The transcription factors regulated by these
pathways, including NFkB (Bayram et al., 2006; Lee et al., 2005a; Takizawa et al., 2003), AP-1
(Donaldson et al., 2003), STAT (Cao et al., 2007), ARE (Li and Nel, 2006) and AhR (Rouse et al., 2008),
have also been studied following PM exposures. Activation of these intracellular pathways and
transcription factors leads to the upregulation of genes responsible for inflammatory, immune and acute
phase responses as well as genes responsible for antioxidant defense and xenobiotic metabolism.
5.1.3. Inflammation
Transcription factor activation in macrophages and epithelial cells stimulates the synthesis and
release of proteins involved in inflammatory and immune responses including cytokines, chemokines,
proteases and eicosanoids (Figure 5-3). These soluble mediators play a role in recruiting inflammatory
cells such as neutrophils, monocytes, mast cells and eosinophils to the lung. Interactions between
macrophages and epithelial cells enhance these responses (Tao and Kobzik, 2002).
Inflammatory cells can also serve as a source of extracellular ROS which, along with soluble
mediators derived from the inflammatory cells, can amplify the inflammatory response. Unchecked
inflammation may cause cellular and tissue injury through the generation of ROS and soluble mediators.
In some cases the oxidative potential of PM is well-correlated with the degree of inflammation (Dick et
al., 2003b), suggesting that the inflammation is a direct consequence of PM-associated ROS. However, in
other cases the oxidative potential of PM is not well-correlated with the degree of inflammation (Beck-
Speier et al., 2005), suggesting that the inflammation is a consequence of the other mechanisms by which
PM activates intracellular signaling pathways.
5.1.4. Epithelial Barrier Function
Epithelial injury can lead to an increase in permeability across the airway epithelial and
alveolar-capillary barriers. Airway and alveolar edema may occur subsequently. Enhanced transport of
December 2008
5-5
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
soluble and insoluble PM components into the circulation may occur under these conditions. Increased
epithelial permeability is also associated with enhanced immune responses to proteins, including
allergens, on the epithelial surface, presumably due to the greater availability of antigens to underlying
immune cells. Soluble mediators derived from inflammatory and lung cells (Chang et al., 2005a) and
peptides released by some nerve cells (Widdicombe and Lee, 2001) can also increase permeability across
epithelial barriers. Edema resulting from nerve cell stimulation is one component of neurogenic
inflammation.
5.1.5. Antioxidant Defenses and Adaptive Responses
Antioxidant defenses and adaptive responses are important modulators of oxidative stress and other
cellular stresses resulting from PM exposure. Antioxidants are present in the epithelial lining fluid at all
levels of the respiratory tract. In addition, they are present in cells of the lung parenchyma and
inflammatory cells found in airways and alveoli. Some antioxidants act directly against oxidant species
(e.g. glutathione, ascorbate, superoxide dismutase) while others act indirectly (e.g. yGCS, glutathione
reductase). Furthermore, some antioxidants (e.g. Phase 2 enzymes HO-1, NQOl GST) are inducible via
activation of the Nrf2-ARE pathway which occurs as an adaptive response to stress (Cho et al., 2006; Li
and Nel, 2006). Antioxidants play an important role in reducing the oxidative potential of those PM
species which directly generate ROS. They also inhibit responses due to generation of intracellular ROS.
Recently a three-tier response to oxidative stress was proposed (Li and Nel, 2006). In this scheme,
mild oxidative stress enhances antioxidant defenses by upregulating Phase 2 and other antioxidant
enzymes (Tier 1). Further increase in oxidative stress induces inflammation (Tier 2) and cell death
(Tier 3). Experimental evidence is supportive of this scheme. Numerous studies have demonstrated that
enhancement of lung and cellular antioxidant defenses inhibits inflammation, cytotoxicity and other
responses following exposure to PM (Ahsan et al., 2005; Bachoual et al., 2007; Bayram et al., 2006;
Chang et al., 2005a; Imrich et al., 2007; Koike et al., 2004; Koike and Kobayashi, 2005; Li et al., 2007;
Liu et al., 2005a; Ramage and Guy, 2004; Rhoden et al., 2004; Steerenberg et al., 2004b; Takizawa et al.,
2003; Tao et al., 2003; Upadhyay et al., 2003; Wan and Diaz-Sanchez, 2006, 2007; Yin et al., 2004b).
Cellular and tissue exposure to xenobiotics carried by PM can lead to induction of Phase 1 and
Phase 2 detoxifying enzymes subsequent to the activation of cell signaling pathways and transcription
factors AhR and ARE, respectively (Rengasamy et al., 2003; Rouse et al., 2008; Zhao et al., 2006a).
December 2008
5-6
DRAFT-DO NOT QUOTE OR CITE
-------
5.1.6. Pulmonary Function
PM exposure may alter pulmonary function by a variety of different mechanisms (Figure 5-4). In
the short term, AHR may ensue due to the influence of inflammatory mediators. In the long-term,
morphological changes ma)' occur, in some cases leading to mucus hypersecretion and airway
remodeling. Activation of irritant receptors and stimulation of the autonomic nervous system (ANS) in the
respiratory tract is another mechanism by which PM exposure may alter pulmonary function
(Section 5.3).
PM Core and
Soluble
Components
Irritant
Receptors
ROS/RNS
Pulmonary
Inflammation
and Injury
Altered
Lung Function
AHR and
Airway
Remodeling
Allergic Asthma
And Other
Allergic
Disorders
Impaired
Host Defense
and Infections
Progression of
Pre-existing
Lung
Disease
DNA Damage
and
Lung
Cancer
Death or Hospitalization for Asthma, Pneumonia, COPD and Lung Cancer
Figure 5-4. Potential pathways for the effects of PM on the respiratory system.
December 2008
5-7
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
5.1.7. Allergic Disorders
PM exposure sometimes leads to the development of allergic immune responses (Figure 5-4).
These responses are mediated by T helper cells, which fall along a continuum from T helper type 1 (Thl)
to T helper type 2 (Th2). Thl responses, characterized by IFN-y, are inflammatory, and in excess can lead
to tissue damage. Alternatively, Th2 responses are characterized by IL-4, IL-5, IL-13, eosinophils, and
IgE, and are associated with allergy and asthma. PM exposure can also lead to the exacerbation of allergic
responses, such as antigen-specific IgE production and AHR.
Due to soluble mediators and immune cell trafficking, pulmonary exposure may result in systemic
immune alterations. Not only do macrophages ingest PM, but they are also antigen presenting cells whose
level of activation dictates costimulation and thus subsequent T cell responses. These cells are highly
mobile, and can transport PM to other sites such as lymph nodes. Dendritic cells (DC) also play a key role
as antigen presenting cells and in modulating T and B cell activity. A cell culture model of the human
epithelial airway wall was used to demonstrate that DC extend processes between epithelial cells through
the tight junctions to collect particles in the lumenal space and to transport them through cytoplasmic
processes between epithelial cells across the epithelium or to transmigrate through the epithelium to take
up particles on the epithelial surface. Furthermore, DC interacted with particle-loaded macrophages on
top of the epithelium and with other DC within or beneath the epithelium to transfer particles (Blank et
al., 2007). In vitro studies also demonstrate that the adjuvant activity of DEP may involve stimulation of
immature monocyte-derived dendritic cells (iMDDC) to undergo maturation by an altered airway
epithelial cell-derived microenvironment (Bleck et al., 2006). Additionally, DEP directly influences the
profile of cytokines secreted by DC and causes a predisposition toward Th2-mediated or allergic
responses (Chan et al., 2006b). Thus PM can negatively affect both innate immunity through effects on
macrophage pathogen handling (see Section 5.1.8) as well as adaptive immunity by altering macrophage
or dendritic cell antigen presenting activity and subsequent T cell responses.
5.1.8. Impaired Lung Defense Mechanisms
PM exposure may impair lung defense mechanisms and result in frequent or persistent infections
(Figure 5-4). Potential targets include mucociliary transport, surfactant function and pathogen clearance.
Pathogen clearance is dependent on the integrity of macrophages and their migration, phagocytosis and
respiratory burst functions. PM-mediated cytotoxicity of macrophages with the concomitant release of
lysosomal contents may affect pathogen clearance and cause damage to nearby cells and tissues.
Intratracheal instillation and cell culture experiments have demonstrated PM-dependent impairment of
December 2008
5-8
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
lung defense mechanisms (Jaspers et al., 2005; Kaan and Hegele, 2003; Long et al., 2005; Moller et al.,
2005; Monn et al.; Roberts et al., 2007; Yin et al., 2004a).
5.1.9. Resolution of Inflammation/Progression of Disease
Resolution of pulmonary inflammation and injury has been demonstrated in many experimental
models using higher than ambient concentrations of PM. Factors contributing to this complex process are
likely to include the uptake and clearance of PM by macrophages, the retention of PM in parenchymal
cells and tissues, pro- and anti-inflammatory soluble mediators, the balance of oxidants/antioxidants and
the presence of pre-existing disease. These factors may also influence the resolution of pulmonary
responses to ambient PM exposures (Figure 5-4). A recent study suggests an important role for retained
particles in the progression of disease. Complexation of endogenous iron by retained particles resulted in
retained particles growing larger overtime. The authors suggested that redox cycling of complexed iron
may be responsible for disease progression (Ghio et al., 2004; Ghio and Cohen, 2005).
5.1.10. Pulmonary DNA Damage
Pulmonary DNA damage can occur primarily or secondarily to PM exposure. Primary effects
include oxidative DNA injury or DNA adduct formation due directly to PM while secondary effects occur
due to PM-mediated inflammation (de Kok et al., 2005; Gabelova et al., 2007b; Gallagher et al., 2003;
Schins and Knaapen, 2007). These responses may lead to chromosomal aberrations or DNA strand
breaks. PM effects on cell cycle arrest, proliferation, apoptosis, and DNA repair mechanisms may also
influence the genotoxic, mutagenic or carcinogenic potential of DNA damage.
5.2. Systemic Inflammation
Pulmonary inflammation can trigger systemic inflammation through the action of cytokines and
other soluble mediators which leave the lung and enter the circulation (Figure 5-5). Cytokines released by
alveolar macrophages can stimulate bone marrow production of leukocytes resulting in an increased
number of total and immature leukocytes in the circulation (van Eeden et al., 2001; van Eeden and Hogg,
2002). They also activate neutrophils and promote their sequestration in microvascular beds (van Eeden et
al., 2001). The time course of these responses varies according to the acute or chronic nature of the PM
exposure (van Eeden et al., 2005).
December 2008
5-9
DRAFT-DO NOT QUOTE OR CITE
-------
Ambient PM
Stroke
Thrombosis
Atherosclerosis
Arrhythmia
Myocardial
Infarction
Myocardial
Ischemia
Pro-
Coagulation
Effects
Autonomic
Nervous System
Death or Hospitalization
for Stroke
Liver
Acute
Phase
Response
Altered
Conduction/
Repolarization
Pulmonary Reflexes
Other Reflexes ?
Endothelial
Cell
Activation/
Dysfunction
Plaque Destabilization
Or Rupture
Systemic
Inflammation/
Oxidative
Stress
Altered
Vasoreactivity
of Coronary
Vessels
Death or
Hospitalization
forThrombo-
Embolic Disease
Pulmonary
Oxidative Stress
and Inflammation
Altered Sympathetic/
Parasympathetic
Tone
Death or Hospitalization for Coronary
Heart Disease or Congestive Heart Failure
Direct Effects ?
T ranslocation/Absorption
of Soluble Components
Figure 5-5. Potential pathways for the effects of PM on the cardiovascular system.
1 Systemic inflammation is seen under conditions of mild pulmonary inflammation - and sometimes
2 under conditions of no measurable pulmonary inflammation - following PM exposure. The time-
3 dependent nature of pulmonary and systemic inflammatory responses may in part explain these findings
December 2008
5-10
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
since biomarkers of inflammation are frequently measured only at one time point. Furthermore, chronic
exposures may lead to adaptive responses. Systemic inflammation is associated with changes in
circulating white blood cells, the acute phase response, pro-coagulation effects, endothelial dysfunction
and the development of atherosclerosis (Figure 5-5). Adverse effects on the cardiovascular and
cerebrovascular systems such as thrombosis, plaque rupture, MI and stroke may result. Systemic
inflammation may affect other organ systems such as the liver or the central nervous systerm.
One recent study demonstrates that alveolar macrophage-derived IL-6 mediated pro-coagulation
effects in mice exposed by intratracheal instillation to 10 jj.g PMi0 (Mutlu et al., 2007). This study
provides a definitive link between lung cytokines and systemic responses in one model system. Whether
this mechanism or others account for the majority of extrapulmonary effects following PM exposure is
not yet clear.
5.2.1. Endothelial Dysfunction and Altered Vasoreactivity
The lumenal surface of blood vessels is lined by endothelial cells which, in addition to providing a
barrier function, are key regulators of vascular homeostasis. Endothelial cells synthesize and release
vasodilators such as nitric oxide and prostacyclin and vasoconstrictors such as endothelin which act on
neighboring smooth muscle cells. Endothelin also stimulates endothelial nitric oxide synthesis through a
feedback mechanism. Inhalation of high concentrations of PM has been reported to increase endothelin
levels in the circulation (Thomson et al., 2005). Endothelin has also been proposed to play a role in
hypoxia-induced MI (Caligiuri et al., 1999). However the role of endothelin in mediating cardiovascular
effects following ambient PM exposures is unclear.
Endothelial dysfunction can arise under conditions of systemic inflammation and/or oxidative
stress. Cytokines activate endothelial cells and upregulate endothelial cell adhesion molecules. They also
promote the sequestration of neutrophils in microvascular beds. Neutrophil sequestration is sometimes
associated with the deposition of myeloperoxidase on endothelial cell surfaces (Nurkiewicz et al., 2006).
ROS-derived from neutrophils, myeloperoxidase, other adhered inflammatory cells and/or other sources
can perturb the balance of vasodilator and vasoconstrictor species produced by endothelial cells.
Oxidative stress can result in decreased synthesis of nitric oxide due to limitation of the redox-sensitive
cofactor tetrahydrobiopterin and in decreased bioavailability of nitric oxide due to reaction with
superoxide. Prostacyclin synthesis is also decreased by oxidative stress. These processes can affect
vasoreactivity, in that blood vessels may be unable to respond to vasoconstrictor stimuli with
compensatory vasodilation.
December 2008
5-11
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Loss of nitric oxide and prostacyclin synthesis due to oxidative stress may have other consequences
since both exert negative influences on platelet and neutrophil activation. While endothelial surfaces
normally are antithrombotic, endothelial dysfunction can contribute to thrombus formation. Furthermore
inflammation and oxidative stress associated with endothelial dysfunction can contribute to the
development or progression of atherosclerosis (van Eeden et al., 2005).
5.2.2. Activation of Coagulation and Acute Phase Response
The primary function of the coagulation cascade is to stop the loss of blood after vascular injury by
forming a fibrin clot. However in some cases, activation of coagulation can promote intravascular
thrombosis (Karoly et al., 2007). It has been proposed that PM air pollution can activate clotting pathways
and enhance the likelihood of an obstructive cardiac ischemic event (e.g. myocardial infarction) or
cerebral event (e.g. stroke) (Seaton et al., 1995).
Coagulation is regulated by intrinsic and extrinsic pathways. The intrinsic pathway occurs
following activation of Factor XII and does not require the addition of an exogenous agent (Mackman,
2005). On the other hand, the extrinsic pathway is an inducible signaling cascade that can be activated by
tissue factor produced in response to inflammation or endothelial injury (Karoly et al., 2007).
In general, platelets, RBC s and endothelial cells are effector cells for inducing a procoagulant state
in the vasculature. Circulating factors may enhance coagulation or promote activation of platelets.
Cytokines formed during tissue damage and inflammation lead to tissue factor induction. Tissue factor is
the initiating stimulus for coagulation following vascular injury or plaque erosion. Complexes of tissue
factor: Factor Vila form on endothelial cell surfaces and play a key role in thrombin generation by
initiating the extrinsic blood coagulation pathway (Gilmour et al., 2005). Thrombin generates fibrin from
fibrinogen and amplifies the intrinsic pathway (Karoly et al., 2007). Tissue factor and thrombin also have
pro-inflammatory actions independent of coagulation functions (Chu, 2005); thus activation of
coagulation may lead to or potentiate inflammation.
The fibrinolytic system opposes these processes by facilitating the removal of a clot. The
fibrinolytic pathway is regulated by the ratio of tissue plasminogen activator (tPA) and plasminogen
activator inhibitor (PAI). Furthermore, the endothelial cell surface has antithrombotic properties due to the
expression of tissue factor pathway inhibitor (TFPI) and thrombomodulin (Mackman, 2005).
Inhibition of the fibrinolytic pathway, along with increased plasma viscosity, plasma fibrinogen and
Factor VII concentrations, contributes to a pro-thrombotic state (Gilmour et al., 2005). In acute lung
injury, vascular cells have enhanced procoagulant activity and impaired fibrinolytic activity (Gilmour et
December 2008
5-12
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
al., 2005). In arterial atherosclerosis, tissue factor expression is increased within plaques. As a result,
spontaneous plaque rupture may trigger intravascular clotting (Karoly et al., 2007).
Acute phase responses also play a role in hemostasis by exerting procoagulant effects. Cytokines
such as IL-6 stimulate the liver to produce acute phase proteins including C-reactive protein (CRP),
fibrinogen and antiproteases (van Eeden et al., 2001).
5.3. Activation of the Autonomic Nervous System by
Pulmonary Reflexes
Chemosensitive receptors, including rapidly activating receptors (RARs) and sensory C-fiber
receptors, are found at all levels of the respiratory tract and are sensitive to irritant particles as well as to
irritant gases (Coleridge and Coleridge, 1994; Widdicombe, 2006). Activation of these vagal afferents
cause central nervous system reflexes resulting in bronchoconstriction, mucus secretion, mucosal
vasodilation, cough, apnea followed by rapid shallow breathing and effects on the cardiovascular system
such as bradycardia and hypotension or hypertension (Coleridge and Coleridge, 1994; Widdicombe and
Lee, 2001; 2006; 2003). Some evidence suggests that cardiovascular responses may be mediated
primarily by the C-fiber receptors (Coleridge and Coleridge, 1994) and that irritants in the lower
respiratory tract cause more pronounced cardiovascular responses than irritants in the upper respiratory
tract (Widdicombe and Lee, 2001).
Early experiments demonstrated that reflexes were mediated by cholinergic parasympathetic
pathways involving the vagus nerve and inhibited by atropine (Grunstein et al., 1977; Nadel et al., 1965a;
Nadel et al., 1965b). However more recent experiments have shown that noncholinergic mechanisms may
also be involved. For example, stimulation of C-fiber receptors can activate local nervous system reflexes.
These local axon pathways are responsible for secretion of neuropeptides and the development of
neurogenic inflammation (Widdicombe and Lee, 2001). It has been proposed that, in some cases,
neurogenic pulmonary responses can switch from their normally protective function to one that
perpetuates pulmonary inflammation (Wong et al., 2003). Differences in respiratory tract innervation
between rodents and humans suggest that C-fiber mediated neurogenic inflammation may be more
important in rodents than in humans (Groneberg et al., 2004; Widdicombe and Lee, 2001; 2003).
However the role of neurogenic inflammation in mediating pulmonary responses in humans is an active
area of investigation.
VR1 receptors represent a subset of neuropeptide and acid-sensitive irritant receptors which are
located on the sensory C-fibers which lie underneath and between lung epithelial cells. In addition, these
December 2008
5-13
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
receptors are found on immune and non-immune airway cells. Recent research interest has focused on the
role played by these receptors in mediating inflammation following exposure to PM (Veronesi and
Oortgiesen, 2001). Exposure to PM has been shown to result in an immediate increase in intracellular
calcium followed by the release of neuropeptides and inflammatory cytokines (Veronesi et al., 1999;
2000). In one study this response was found to be due to an intrinsic property of the particle core and was
not metal-dependent (Oortgiesen et al., 2000), while in another study electrostatic charge was found to
activate VR1 receptors (Veronesi et al., 2003). A recent study demonstrated PM-mediated activation of
VR1 receptors which results in increases in intracellular calcium and apoptosis in epithelial cells
(Agopyan et al., 2003).
At this time, it is not clear how activation of the autonomic nervous system by pulmonary reflexes
may contribute to the kinds of altered conduction and/or repolarization properties of the heart which may
be linked to arrhythmias (Figure 5-5). Pulmonary reflexes, as they are currently understood, initially lead
to increases in parasympathetic tone. However decreased heart rate variability appears to be reflective of
decreased parasympathetic tone and/or increased sympathetic tone. Thus activation of the autonomic
nervous system by mechanisms other than pulmonary reflexes seems likely in response to PM. Very little
is known about these putative alternative mechanisms although some new studies have focused on the
role of CNS centers in regulating pulmonary and cardiovascular functions. Furthermore the effects of
pre-existing alterations in the autonomic nervous system on PM responses are not understood.
5.4. Translocation of Ultrafine PM or Soluble PM
Components
Ultrafine PM is small enough to be taken up by cells through endocytosis and transcytosis.
Localization of ultrafine PM in macrophage mitochondria has been demonstrated by electron microscopy
(Li et al., 2003). Whether particles are capable of crossing the epithelial barrier and reaching capillary
endothelial cells or the circulation is in question (Figure 5-5). To date, the evidence for ultrafine or other
PM size fractions accessing the circulation by traversing this barrier is not convincing (see Chapter 4).
However, macrophage-associated particles may be transported to the lymph nodes and gastrointestinal
system and gain access to the circulation by an alternate mechanism.
Soluble components from all size fractions of PM have the potential to translocate across the
alveolar-capillary barrier into the circulation (Figure 5-5). Possible mechanisms involved in translocation
include paracellular pathways and metal transporters. Wallenborn et al. (2007) demonstrated the rapid
appearance of water-soluble metals in the blood, heart and liver following intratracheal instillation of oil
December 2008
5-14
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
combustion PM in rats (see Chapter 4). Similarly, Gilmour et al. (2006b) demonstrated the rapid
appearance of zinc in the plasma of rats intratracheally instilled with zinc sulfate. Soluble zinc was also
associated with cardiac effects following intratracheal instillation of rats with zinc-containing PM
(Kodavanti et al., 2008). Inhalation studies involving concentrations of PM relevant to ambient exposures
have not yet demonstrated the translocation of soluble components.
Interaction of circulating PM or soluble PM components with vascular endothelial cells, platelets,
and other leukocytes is a potential mechanism underlying the cardiovascular effects of inhaled PM. A role
for PM-derived ROS and/or cellular-derived ROS has been proposed. Furthermore, soluble metals which
do not redox-cycle may activate cell signaling pathways without the generation of ROS. In this way PM
may promote adverse cardiovascular effects such as endothelial dysfunction, atherosclerosis and
thrombosis. Circulating PM or soluble PM components also have the potential to impact other organ
systems.
5.5. Disease of the Cardiovascular and Other Organ
Systems
As discussed above, deposition of PM in the lung may lead not only to pulmonary disease but also
to diseases of other systems (Figure 5-5). In the cardiovascular system, myocardial ischemia and MI may
occur as a result of the above proposed effects on atherosclerosis, plaque instability, thrombosis, plaque
rupture and/or altered vasoreactivity of coronary vessels. Myocardial ischemia and MI may alter the
conduction and depolarization properties of the heart and lead to arrhythmic events. In addition,
thrombosis may lead to stroke and/or thromboembolic disease. At this point, it seems that many of these
processes are interlinked and that responses to ambient PM exposures may involve multiple mechanisms
simultaneously with some variability depending on PM composition.
5.6. Results of New Inhalation Studies which Contribute to
Modes of Action
Prior to this review, much of the evidence for the proposed modes of action was obtained from
animal studies involving intratracheal instillation or inhalation of high concentrations of PM and from cell
culture experiments. In many cases, the types of PM used were of questionable relevance to ambient
exposures (i.e. ROFA, metals, ambient PM collected on filters). Since then, many inhalation studies have
been conducted using CAPs, combustion-derived PM and BC, generally using concentrations of PM
December 2008
5-15
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
lower than 1 mg/m3. Much of this research has been conducted in animal models of disease. These key
new studies, described in detail in Chapters 6 and 7, add to the understanding of modes of action which
are relevant to ambient PM exposure. A compilation of pertinent results is found below.
¦ Mild pulmonary inflammation in response to short-term exposures to CAPs and combustion-
derived PM.
¦ Pulmonary, cardiovascular and systemic oxidative stress in response to CAPs, combustion-
derived PM and "inert particles."
¦ Antioxidant intervention which ameliorates PM effects on oxidative stress, allergic responses,
and airway hyperresponsiveness.
¦ Altered lung function including respiratory frequency and airway responsiveness following
short-term exposures to CAPs and combustion-derived PM.
¦ Allergic sensitization and exacerbation of allergic responses, in response to CAPs and
combustion-derived PM.
¦ Increased susceptibility to respiratory infection following exposure to diesel PM.
¦ Effects on nasal epithelial mucosubstances, airway morphology and airway mucosubstances
in chronic studies using roadside air and woodsmoke.
¦ Effects on lung development in chronic studies using roadside air.
¦ A role for irritant receptors in activating local neural and CNS reflexes through
parasympathetic pathways following short term exposure to CAPs and diesel PM.
¦ A role for TRPV1 irritant receptors in mediating lung and heart oxidative stress through
increased parasympathetic and sympathetic activity in response to CAPs.
¦ Altered heart rate variability in response to CAPs and combustion-derived PM.
¦ Arrhymthmic events in response to CAPs.
¦ Decreased cardiac contractility following short-term exposure to PM
¦ Enhanced myocardial ischemia in response to CAPs.
¦ Endothelial dysfunction and altered vascular reactivity in response to short- and long-term
exposure to CAPs.
¦ Increased levels of blood coagulation factors following short-term PM exposures.
December 2008
5-16
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
¦ Progression of atherosclerosis and induction of tissue factor in aortic plaques following long-
term exposure to CAPs.
¦ CNS responses following short- and long-term exposures to CAPs.
¦ DNA adducts in lung and liver and heritable DNA mutations in germ line cells in long-term
studies involving roadside air.
However new studies do not provide a complete picture of the biological pathways involved in
mediating PM effects. In fact, there is a lack of information regarding the time-dependence of many of the
responses which makes it difficult to understand the underlying biological mechanisms. Other existing
gaps in knowledge include:
¦ Effects of ambient PM exposures on epithelial barrier function in the lung.
¦ The putative modulation of neural reflexes involving vagal parasympathetic pathways by pre-
existing disease or other factors.
¦ The putative role of other neural reflexes besides vagal parasympathetic pathways.
¦ The putative role of endothelin in altering vasomotor tone following PM exposure.
¦ The putative translocation of PM or soluble components across the epithelial barrier of the
lung into the circulation.
¦ The putative translocation of PM from olfactory epithelium to the olfactory bulb and other
brain regions.
December 2008
5-17
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Chapter 6. Integrated Health Effects of
Short-Term PM Exposure
6.1. Introduction
This chapter summarizes, reviews and integrates the evidence of relationships between short-term
exposures to PM and a variety of health-related endpoints. Cardiovascular and respiratory health effects
of short-term exposure to various size fractions and sources of PM have been examined in an expansive
number of epidemiologic, human clinical and toxicological studies. In addition, there is a large body of
literature evaluating the relationship between mortality and short-term exposure to PM. The association
between PM exposure and central nervous system function has also been assessed, although far fewer
studies are available. The research approaches used to evaluate health effects of PM exposure are
described in Section 1.5 along with advantages and limitations of the various study types. Chapter 5
provides an overview of the potential pathophysiological pathways and modes of action underlying the
PM-induced health effects observed in animal and human studies. Evidence from the scientific literature
of specific cardiovascular and systemic effects, respiratory effects, and central nervous system effects
associated with exposure to PM are presented in Sections 6.2, 6.3, and 6.4, respectively. More detailed
descriptions of each study evaluated for this assessment are presented in Annexes C, D, and E. Evidence
of associations between short-term exposure to PM and mortality are described in Section 6.5, along with
causal determinations of mortality by PM metric (Section 6.5.3). The chapter concludes with a
preliminary evaluation of PM-induced health effects attributable to specific constituents or sources
(Section 6.6).
Findings for cardiovascular and respiratory effects are presented by specific endpoint or measure of
effect, leading from more subtle health outcome measures (e.g., heart rate variability [HRV]) to the more
severe, such as hospitalization for cardiovascular disease. Conclusions from the 2004 PM AQCD are
briefly summarized at the beginning of each section, and the evaluation of evidence from recent studies
builds upon what was available during the previous review. For each health outcome, results are
summarized for studies from the specific scientific discipline, i.e., epidemiologic, human clinical, and
toxicological studies. The sections conclude with summaries of the evidence on the various health
outcomes and integration of the findings that leads to conclusions regarding causality based upon the
framework described in Chapter 1. Determination of causality is made for the overall health effect
category, such as cardiovascular effects, with coherence, consistency and plausibility being based upon
December 2008
6-1
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
the evidence from across disciplines and also across the suite of related health outcomes. In these
summary sections for cardiovascular and respiratory morbidity, the evidence is summarized and
independent conclusions drawn for relationships with PM10, PM10_2.5, PM2 .5, and ultrafine particles
(Sections 6.2.11 and 6.3.9, respectively). Evidence of central nervous system effects is also divided by
scientific discipline (Section 6.4), however, the lack of data does not allow for informative summaries of
effect by PM metric.
6.1.1. Methodological Considerations
6.1.1.1. Epidemiologic Studies
Although the multicity studies analyzed in the 2004 PM AQCD reported an association between
short-term exposure to PMi0 and mortality, these studies also found large spatial heterogeneity in
city-specific excess risk estimates. Similar results have also been observed for studies that analyzed other
health outcomes (e.g., respiratory and cardiovascular hospital admissions and mortality associated with
long-term exposure). The reasons for such variation in effects estimates were not well understood, but
factors likely contributing to the apparent heterogeneity between cities have been identified, including,
but not limited to geographic differences in: (1) air pollution mixtures; (2) composition and sources of
ambient PM; and (3) personal and sociodemographic factors potentially affecting PM exposure (e.g., air
conditioning use). Overall, studies published at that time provided insufficient information to adequately
examine the factors that contribute to the heterogeneity observed in effect estimates.
Copollutants have generally been considered as potential confounders or effect modifiers of the
health effect attributed to a specific air pollutant. The 2004 PM AQCD conducted a detailed examination
of confounding and effect modification of mortality and morbidity effect estimates due to copollutants.
The overall evidence tended to support the conclusion that short-term exposure to ambient PMi0 and
PM2 5 is most clearly associated with mortality and morbidity effects, acting either alone or in
combination with other covarying gaseous pollutants, with more limited support for PMi0.2.5. The results
from multicity studies suggested that PM-mortality/morbidity associations were generally robust to the
inclusion of copollutants. This draft ISA presents results from epidemiologic studies that report
multipollutant model results.
Epidemiologic studies have used a variety of approaches to control for weather effects
(i.e., meteorological variables) to disentangle the true effect due to PM. Various studies were identified
that appear to demonstrate increased PM-related mortality/morbidity risks beyond those attributed to
weather influences alone. However, similar to the issues surrounding models developed to control for
December 2008
6-2
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
potential confounding, a clear consensus was not reached as to what constitutes an appropriate or
adequate model to control for possible weather contributions to the mortality/morbidity effects attributed
to PM exposure. In addition, it was unclear on how best to characterize possible joint (interactive) effects
of weather and ambient PM or other air pollutants on mortality.
To date, epidemiologic studies have used various lag structures to examine the association between
air pollution and health effects. The maximum effect sizes for mortality due to short-term exposure to PM
have often been reported for 0-1 day lags, but evidence is beginning to suggest that more consideration
should be given to lags of several days (e.g., distributed lags). It has also been hypothesized that different
PM size or chemical components may produce effects by different mechanisms manifested at different
lags, but limited data was available to substantiate these claims.
The majority of studies analyzed in the 2004 PM AQCD that examined the association between
short-term exposure to PM and mortality/morbidity used time-series analyses. Although the time-series
study design allows for a detailed analysis of the affect of daily fluctuations in PM levels on
mortality/morbidity, complex modeling is required to control for seasonal variation, time trends, and slow
time varying confounders. To date, a clear consensus as to the extent of modeling required to accurately
measure PM-mortality/morbidity effects has not been reached. The case-crossover study design was
developed as an alternative to using complex models to control for confounders. Some of the control
selection procedures that have been used in case-crossover studies include 'unidirectional' control
selection (i.e., control periods all before the case/hazard period), 'bi-directional' control selection
(i.e., control periods equally spaced before and after the case period [e.g. 7 days]), and 'time-stratified'
control selection (i.e., control periods in the same calendar month). Although the 2004 PM AQCD did not
conclude, which control/referent selection procedure should be used, a recent comprehensive review by
Janes et al. (2005) examined each of the control/referent selection strategies used in case-crossover
studies of air pollution, and recommended the use of the time-stratified control selection procedure for use
in all future case-crossover air pollution studies (Janes et al., 2007).
Although the majority of studies only analyzed the association between short-term exposure to
ambient levels of PM and mortality, some multi- and single-city studies also conducted analyses to
examine the presence of a threshold. Overall, the results from large multicity studies suggest that strong
evidence does not exist for a clear threshold for PM mortality/morbidity effects. However, some
single-city studies suggest a hint of a threshold, but not in a statistically clear manner. As a result, the PM
AQCD concluded that more data is needed to answer the question of whether a threshold exists, but the
use of linear PM effect models appears to be appropriate.
Lastly, the 2004 PM AQCD examined the role of measurement error in time-series epidemiologic
studies using a simulation study and mathematical analyses. The simulation study indicated that "transfer
December 2008
6-3
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
of effects" (i.e., the effects of one pollutant are inappropriately attributed to another in multi-pollutant
models) occurred when the correlation between the true predictor and the confounder was very high (r >
0.90) with the true predictor having moderate error (o >0.5) and the confounder having no error, but
transfer of effects lessened as the confounder became subject to error (U.S. EPA, 2006b). The
mathematical analyses found that only under unusual circumstances (i.e., true predictors having high
positive or negative correlation; substantial measurement error; or extremely negatively correlated
measurement errors) did weak predictors appear stronger than true predictors. It was concluded that some
of the conditions reported that could potentially lead to the transfer of effects have not been observed in
actual air pollution data.
With regard to issues surrounding the design of air pollution epidemiologic studies the 2004 PM
AQCD concluded: (1) heterogeneity exists between cities, which can primarily be attributed to personal
and sociodemographic factors, and PM sources/composition; (2) although evidence supports the
conclusion that short-term exposure to PMi0 and PM2 5 (and to a lesser extent PMi0_2.5) alone or in
combination with gaseous copollutants are associated with mortality/morbidity, it still remains unclear
which statistical methods should be used to identify potential confounding, and insufficient information is
available to examine the effect modification of PM by copollutants; (3) a consensus has not been reached
as to the best approach to control for meteorological effects or the joint (interactive effects) of weather
and air pollutants on mortality/morbidity; (4) although the previous review did not conclude which
control/referent selection procedure should be used in the case-crossover study design, the time referent
approach has gained increasing credibility; (5) the maximum mortality/morbidity effect size attributed to
PM has been observed for 0-1 day lags with lags of multiple days possibly showing the greatest effects,
and limited data exists to examine the lags associated with PM size and chemical components; (6) the
results from single- and multicity studies suggest that it seems appropriate to use a linear threshold during
the examination of mortality effects attributed to PM exposure; and (7) measurement error, specifically
the "transfer of effects," has only been observed in simulations and mathematical analyses under unusual
circumstances, which some have not been reported in real world scenarios.
6.1.1.2. Experimental Studies
CAPs
The Methodological Considerations Section of Chapter 7 in the 2004 PM AQCD provides a brief
discussion of particle concentrators used in concentrated ambient particles (CAPs) studies. Particle
concentrators enable human subjects, animals, or cell culture systems to be exposed to atmospheric PM at
concentrations much greater than that observed under ambient conditions. As ambient PM is just one
December 2008
6-4
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
component of a complex mixture that interacts with gases and other aerosols, CAPs systems provide a
method of exposing subjects to the particle phase. Gases (such as 03 and S02) are not concentrated nor
are thoracic coarse particles (except for the coarse particle concentrator) and only certain systems are
capable of concentrating ultrafine PM. In ultrafine CAPs systems, increased number fraction of organic
carbon and PAHs, along with decreased relative percentage of EC particles have been reported in
concentrated PM compared to ambient PM (Su et al., 2006). These data suggest that for ultrafine
concentrators, the CAPs do not accurately reflect atmospheric ultrafine composition. There are several
instrument systems used to concentrate ambient PM in controlled human or animal exposure studies
(Gordon et al., 1999; Maciejczyk et al., 2005; Sioutas et al., 1995; Sioutas et al., 1997; Sioutas et al.,
1999).
The system developed by Sioutas et al. (1995) and Sioutas et al. (1997) was based on a series of
three virtual impactors, and concentrates ambient particles in the aerodynamic size range from 0.15 |_im to
2.5 |_im by a factor of 25-30. Ultrafine PM cannot be effectively concentrated by the system. Sioutas et al.
(1997) evaluated the system and reported that the concentration factors for particles in different size
ranges were different (23.3 ± 0.7 for 0.15-0.25 (im, and 26.9 ± 1.0 for 1.0-2.5 (im) as a result of the
different collection efficiencies for particles with different sizes. In addition, Sioutas et al. (1997)
observed a similar concentration factor for nitrate, which is similar to sulfate and fine PM mass. However,
the sampling artifact of nitrate on Teflon filters was not examined during the evaluation study, and
therefore, the loss of semi-volatile PM components during concentration was not clear. Furthermore, the
increase in PM concentration could potentially increase particle losses through coagulation and therefore
change particle size distribution; and the large pressure drop through the virtual impactors might change
the surface properties of a particle constituent without changing its bulk profile.
The method developed by Gordon et al. (1999) are based on a cyclone, and can concentrate
ambient PM in the aerodynamic size range of 0.5-2.5 |_im by a factor of 10. Similar to Sioutas et al.
(1995), the system can not effectively concentrate ultra-fine PM. Concentration factors for PM with
different sizes, and therefore the particle size distribution for CAPs, were observed to be a function of the
system tuning conditions. Cautions need to be exercised when comparing CAPs produced under different
instrument conditions.
An ultrafine PM concentrating system was developed by Sioutas et al. (1999) and Maciejczyk et al.
(2005). The system was based on the Condensation Particle Counter (CPC) theory to let a particle grow
until the inertia force could effectively function on the particle, so that the grown particles could be
concentrated by the two systems mentioned above. After concentration, excess water on particles was
removed and particles shrunk to their original sizes. Sioutas et al. (1999) reported that the system could
effectively concentrate ultrafine particles, and no coagulation and nitrate evaporation was observed.
December 2008
6-5
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
However, the results presented by Sioutas et al. (1999) suggested that the particle size distribution of
ultrafine PM could be changed after concentration under ambient conditions. Sioutas et al. (1999) also
suggested that the PM toxicity might be modified by the system due to the re-distribution of water soluble
components within a particle. In addition, it is possible that some of the physical properties, such as PM
surface properties and particle shape, could also be modified by the system.
A coarse particle concentrator has recently been developed by the civil engineering department of
the University of Southern California that consists of virtual impactors in parallel and can enrich ambient
PMio.2.5 concentrations by a factor of 8-30 (Chang et al., 2002).
Human Clinical Study Advantages and Limitations
Controlled human exposure studies evaluate the effects of exposures to a variety of pollutants in a
highly controlled laboratory setting. Also referred to as human clinical studies, these experiments allow
investigators to expose subjects to fixed concentrations of air pollutants under carefully regulated
environmental conditions and activity levels. Controlled human exposures to particulate matter typically
involve exposing subjects either at rest or while engaged in intermittent exercise in a whole-body
exposure chamber, although mouthpiece and facemask systems can also be used. A variety of different
types of particles are used in these studies including ambient outdoor particles, concentrated ambient
particles (CAPs), diesel exhaust from a diesel engine, wood smoke generated in a wood stove, laboratory
generated surrogate particles (e.g., elemental carbon or zinc oxide), or particles collected on a filter,
resuspended in saline, and administered either through instillation or inhalation (aerosolized and delivered
using a nebulizer). The recovery of particles on filters is variable and some components, such as organics,
may be too volatile to be collected. Exposures to processed emission artificially generated particles may
provide important information on the health effects of particulate matter, but are not truly representative
of ambient air pollution particles. The direct exposure of humans to ambient air pollution particles may be
complicated by factors that cannot be controlled such as co-exposures to other air pollutants (e.g., 03,
S02, and N02). In concentrating ambient particulates, gaseous co-pollutants are not proportionately
concentrated and interactions between PM and the co-pollutants cannot be investigated unless the latter
are re-introduced. These limitations as well as daily variability in concentration and composition can
make it difficult to compare the results of human clinical studies employing particles from different
sources.
Human clinical studies are valuable in characterizing the associations observed in epidemiologic
studies between exposure to particulate matter and a given health endpoint. In some instances, these
studies can also be used to characterize concentration-response relationships at pollutant concentrations
relevant to ambient conditions. Controlled human exposures are typically conducted using a randomized
December 2008
6-6
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
crossover design with subjects exposed both to PM and a clean air control. In this way, subjects serve as
their own controls, effectively controlling for many potential confounders. However, human clinical
studies are limited by a number of factors including a small sample size and short exposure time. The
repetitive nature of ambient PM exposures may lead to cumulative health effects, but this type of
exposure is not practical to replicate in a laboratory setting. In addition, although subjects do serve as
their own controls, personal exposure to pollutants in the hours and days preceding the controlled
exposures may vary significantly between and within individuals. Finally, human clinical studies require
investigators to adhere to stringent health criteria for a subject to be included in the study, and therefore
the results cannot necessarily be generalized to an entire population. Although some human clinical
studies have included health comprised individuals such as asthmatics or individuals with COPD or
coronary artery disease, these individuals must also be relatively healthy and do not represent the most
sensitive individuals in the population.
Selection Criteria for Key Toxicological Studies
The majority of toxicological studies highlighted in the text were selected if the exposure route was
inhalation and the exposure concentration was within 3 orders of magnitude of ambient PM
concentrations. Most of these studies utilized ambient particles (including CAPs, PM collected on filters,
or PM obtained from soil or road surfaces) or diesel exhaust; the remainder of studies utilized gasoline
emissions, wood smoke, residual oil fly ash, carbon black, or titanium dioxide (Ti02). A few studies were
included that employed intratracheal instillation techniques, mainly for thoracic coarse PM studies in
rodents, new emerging areas of investigation (e.g., vasoreactivity, relative toxicity of different size
fractions, etc.), or in attempts of elucidating specific pathways or mechanisms of response. Only a select
number of in vitro studies that examined pulmonary toxicity were included and studies that looked at
direct particle effects on cardiac or vascular cells were excluded, as their relevance based on the current
state of knowledge regarding particle translocation is unclear. All toxicological studies examined for
inclusion in the PM ISA are included in Appendix D.
Apolipoprotein E knockout mouse
Atherosclerosis and related pathways has been studied primarily in the Apolipoprotein E (ApoE)
knockout mouse. Developed by Nobuyo Maeda's group in 1992 (Piedrahita et al., 1992; Zhang et al.,
1992), the ApoE" " mouse and related models have become the workhorse of atherosclerosis research over
the past 15 years. The ApoE molecule is involved in the clearance of fats and cholesterol. When ApoE (or
the LDL receptor) is deleted from the genome, mice develop severely elevated lipid and cholesterol
profiles; ApoE" " mice on a high-fat ("Western") diet exhibit cholesterol levels exceeding 1000 mg/dl
December 2008
6-7
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
(normal is -150 mg/dl) (Huber et al., 1999; Moore et al., 2005). As a result, the lipid uptake into the
vasculature is increased and the atherosclerotic process is dramatically hastened. Furthermore, the LDLs
in ApoE" " mice are highly susceptible to oxidation (Hayek et al., 1994), which may be a crucial event in
the air pollution-mediated vascular changes.
6.2. Cardiovascular and Systemic Effects
6.2.1. Heart Rate and Heart Rate Variability
Heart rate (HR), HRV, and BP are all regulated, in part, by the sympathetic and parasympathetic
nervous systems. Changes in one or more may increase the risk of cardiovascular events (e.g.
arrhythmias, myocardial infarction, etc.). Decreases in HRV have been associated with cardiovascular
mortality/morbidity in older adults and those with significant heart disease (TFESCNASPE, 1996).
HRV is measured using electrocardiograms (ECG) and can be analyzed in the time domain (e.g.
standard deviation of all NN intervals [SDNN], square root of the mean squared successive NN interval
differences [r-MSSD]), and/or the frequency domain measured by power spectral analysis (e.g. high
frequency [HF], low frequency [LF], ratio of LF to HF [LFHFR]). SDNN generally reflects the overall
modulation of HR by the autonomic nervous system, whereas r-MSSD generally reflects parasympathetic
activity and high frequency variations in HR. Thus, r-MSSD is generally well correlated with HF, which
also reflects the parasympathetic modulation of HR. LF primarily reflects sympathetic modulation of HR,
but may also estimate the contribution of both sympathetic and parasympathetic influences on HRV. Thus
LFHFR is thought to estimate the ratio of sympathetic influences on HRto parasympathetic influences.
While HRV is commonly described as being a reflection of vagal and adrenergic input to the heart,
there is clearly a more complex phenomenon reflected in HRV parameters. Rowan et al. (2007) provide a
review of HRV and its use and interpretation with respect to air pollution studies. To summarize, HRV
indices are excellent measures of extrapulmonary effects from inhaled pollutants, but the characterization
of the acute, reversible responses as being either parasympathetic or sympathetic in origin, much less
predictive of some adverse outcomes such as ventricular arrhythmia, is relatively unsupported by the
clinical literature. This is consistent with the 2004 PM AQCD which stated that there is inherent
variability in the minute-to-minute spectral measurements, but long-term HRV measures demonstrate
excellent day-to-day reproducibility (U.S. EPA, 2004).
The 2004 PM AQCD presented limited evidence of PM-induced changes in HRV. However,
findings from epidemiologic, human clinical and toxicological studies were seemingly contradictory, with
December 2008
6-8
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
reports of both decreases and increases in HRV following PM exposure. Recent epidemiologic studies
have demonstrated a more consistent decrease in HRV (SDNN and r-MSSD), which is supported by
several human clinical studies published since 2003. In these studies, decreases in HRV were observed in
HRV among healthy adults following short-term exposures to PM2 5 and PMi0_2.5 CAPs. It is interesting to
note that these effects were not observed in adults with asthma or COPD. The effect of PM on HRV
observed in animal toxicological studies continues to vary greatly, which may be due in part to strain
differences in baseline HRV.
6.2.1.1. Epidemiologic Studies
The 2004 PM AQCD reviewed several studies of PM exposure and HR or HRV and decribed
discrepant findings across studies (U.S. EPA, 2004). Several studies have investigated the association
between acute changes in multiple HRV parameters and ambient air pollutant concentrations in the U.S.,
Canada, Europe, Mexico, and Asia. Features and results of these studies are presented in Table 6-1, and
are summarized below.
In a multicity study, Liao and colleagues (2004) used data from the fourth cohort evaluation of the
Atherosclerosis Risk in Communities (ARIC) Study (1996-1998) (Liao et al., 2004). The 6784 subjects
were 45-64 years of age and lived in either Washington County, MD, Forsyth County, NC, or the suburbs
of Minneapolis, MN. The mean PMi0 concentration during the study is shown in Table 6-1. At each HRV
measurement session, each subject rested comfortably for 15 mins in the supine position in a quiet,
semi-dark room, with a constant temperature of 24°C. Then, resting, supine, 5-min beat-to-beat RR-
interval data were collected. All measurements were made between 8:30 a.m. and 12:30 p.m. Liao et al.
(2004) used linear regression models, adjusting for multiple covariates (i.e. age, ethnicity, gender,
education, smoking, body mass index, cardiovascular medications, presence of coronary heart disease,
diabetes, hypertension, HR, humidity, temperature, and season), to examine the change in HRV associated
with PM10, 03, S02, CO, and N02 concentrations in the 1-3 days prior to ECG measurement. Among all
subjects, each 11.5 (ig/m3 increase in mean daily PMi0 concentration 1 day before the ECG measurement
was associated with a 0.06 ms2 decrease in log-transformed HF (95% CI: -0.10 to -0.02) and a 1.03 ms
decrease in SDNN (95% CI: -1.64 to -0.42). A smaller non-significant decrease was also observed for log
transformed LF. These HRV changes were larger among hypertensive subjects (Liao et al., 2004)
suggesting this may be a group particularly susceptible to the autonomic effects of PM.
Timonen et al. (2006) conducted a multicity panel study of n = 131 elderly subjects with stable
coronary heart disease who lived in 3 European cities (Amsterdam, Netherlands; Erfurt, Germany; or
Helsinki, Finland) (Timonen et al., 2006). They collected ECGs biweekly for six months in each subject.
This analysis, done as part of the ULTRA Study, examined changes in HRV (resting, paced breathing,
December 2008
6-9
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
supine, and five min beat-to-beat NN intervals) associated with changes in fixed monitor particulate
concentrations (PM25, PM10.2.5) with an emphasis on ultrafine particle counts (UFP; 0.01-0.1 (.im
particles) and counts of accumulation mode particles (ACP; 0.1-1.0 |_im particles). The mean city-specific
particulate concentrations are shown in Table 6-1. Mixed models adjusting for time trend, weekday,
humidity, barometric pressure, and temperature were first fit to estimate the change in HRV associated
with particulate (UFP, ACP, PM2 5, and PMi0.2 5) concentrations on the same and previous four days in
each city. Then, in pooled analyses, the most consistent results were for LFHFR. During paced breathing,
each 10,000 particles/cm3 increase in two day lagged UFP was associated with a 13.5% decrease in
LFHFR (95% CI: -20.1 to -7.1). Each 1000 particles/cm3 increase in 1 day lagged ACP was associated
with a 7.8% decrease in LFHFR (95% CI: -13.0 to -0.2). Although not statistically significant, each
10 (ig/m3 increase in 2 day lagged mean PMi0_25 concentration was associated with a 3.3% decrease in
LFHFR (95% CI: -12.7 to 6.1). For PM2 5, however, results were not consistent across cities, and thus a
pooled estimate was not appropriate. PM2 5 was associated with decreased HF power and increased
LFHFR in Helsinki, increased HF power and decreased LFHFR in Erfurt, and not associated with any
HRV metric in Amsterdam. The authors state that these contrasting city-specific PM2 5 findings do not
clearly support a PM/HRV association, but suggest that effects may be dependent on PM sources and
subject characteristics in each city (Timonen et al., 2006).
The association between HRV and short-term increases in PM was also examined in single-city
studies (Table 6-1). Among U.S. and Canadian cities, increases in PM2 5 were generally associated with
decreased SDNN (Adar et al., 2007a; Chahine et al., 2007; Park et al., 2005b; Pope et al., 2004a;
Schwartz et al., 2005b) and/or decreased HF power (Adar et al., 2007a; Chahine et al., 2007; Park et al.,
2005b; Park et al., 2006b; Park et al., 2008a; Schwartz et al., 2005a), but not in all studies. Two studies
reported increased SDNN associated with PM2 5 concentrations (Riediker et al., 2004b; Wheeler et al.,
2006a). Yeatts et al. (2007) also reported increased r-MSSD, SDANN5 (standard deviation of the average
of normal to normal intervals in all 5-min intervals in a 24-h period), and SDNN24HR (standard deviation
of the average of all normal to normal intervals in a 24-h period), and HF power associated with increased
PM2 5 concentrations (Yeatts et al., 2007).
Lipsett et al. (2006) reported significantly decreased SDNN associated with increases in 2- and 6-h
mean PMi0 and PMi0.2 5 concentrations (Lipsett et al., 2006). Similarly, Yeatts et al. (2007) reported
decreased r-MSSD, SDNN24HR, SDANN5, ASDNN5 (mean of the standard deviation in all 5-min
segments of a 24 h recording), proportion of NN intervals < 50 ms apart (pNN50; 7-min and 24-h), and
HF power ssociated with increased PMi0.2.5 concentration.
Of those studies examining HRV associations with particle counts (Adar et al., 2007a; Park et al.,
2005b), only Adar et al. (2007a) found clear evidence of such effects. Decreased HRV was also associated
December 2008
6-10
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
with increases in ambient mean sulfate concentration (Luttmann-Gibson et al., 2006; Park et al., 2008a)
ambient mean BC concentration, (Park et al., 2005b; Park et al., 2008a; Schwartz et al., 2005b), and
traffic generated particles/pollution (Adar et al., 2007a). Riediker et al. (2004b) reported increased HRV
associated with traffic generated pollution (2004a; Riediker et al., 2004b).
Studies in Asia and Mexico have also reported decreased HRV associated with increases in PM
concentration (Chan et al., 2004; Chuang et al., 2005a; 2007b; Holguin et al., 2003; Romieu et al., 2005;
Vallejo et al., 2006). However, Riojas-Rodriguez et al. (2006) reported significantly decreased LF and HF
power associated with each 1 ppm increase in CO concentration, but only small non-significant decreases
associated with PM2 5.
HRV studies investigated lagged pollutant concentrations from 2 h to 5 days before ECG
measurement, reporting effects associated with mean pollutant concentrations lagged as short as 1-2 h,
and more consistently lagged 24-48 h. Taken together, these international and U.S./Canadian studies show
decreases in HRV associated with PM in most studies that use SDNN (14 of 17), and most (12 of 19)
studies that use r-MSSD or HF power. However, these proportions may be inflated by publication bias
(i.e., studies showing little or no effects are not submitted for publication). Studies of HRV and PM
concentration are summarized in Table 6-1.
HRV Studies Investigating Specific Mechanisms
Panel studies investigating PM/HRV associations have also been useful in investigating potential
mechanistic pathways by which PM may elicit a cardiovascular response. Romieu et al. (2005)
hypothesized that the omega-3 fatty acids in fish oil supplements would mitigate the adverse effects of
acute PM exposure on HRV. In a randomized controlled trial of n = 50 residents of a Mexico City nursing
home (aged >60 years), subjects were randomized to either 2 g/day of fish oil or 2 g/day of soy oil. They
used random-effects regression models to estimate the change in HRV associated with mean PM2 5
concentration in the pre-supplementation and supplementation phases. In the group receiving the fish oil
supplement, each 8 (ig/m3 increase in 24-h mean total PM2 5 exposure (weighted average of indoor and
outdoor PM2 5 based on time activity diaries) was associated with a 54% reduction (95% CI: -72% to
-24%) in log transformed HF in the pre-supplementation phase. However, in the supplementation phase of
the trial, each 8 (ig/m3 increase in 24-h mean total PM2 5 concentration was associated with only a 7%
reduction in log transformed HF (95% CI: -20% to 7%). Decreases in other HRV parameters associated
with PM2 5 were also muted in the supplementation phase. In the group receiving the soy oil supplement,
the reduction in HF was also smaller in magnitude during the supplementation phase. However, among
those receiving the soy oil supplement, the differences between the pre-supplementation PM2 5/HF change
and the supplementation PM2 5/HF change were smaller compared to those receiving the fish oil, and were
December 2008
6-11
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
not statistically significant (Romieu et al., 2005). Romieu et al. (2008) also report that omega-3
polyunsaturated fatty acids appear to modulate the adverse effect of PM2 5 based on measured biomarkers
of oxidative response (see Section 6.2.9.1.).
By studying effect modification of the PM/HRV association by genetic polymorphisms, subject
characteristics, and chronic lead exposure, a series of analyses using data from the Normative Aging
Study has also provided mechanistic insights into the PM/HRV association (Chahine et al., 2007; Park et
al., 2005b; 2006b; 2008a; Schwartz et al., 2005a). Park et al. (2005b) studied the association between
short-term increases in ambient air pollution and changes in HRV using n = 497 males from the
Normative Aging Study (mean age 72.7 years) living in the Boston metropolitan area. Subjects had ECG
measurements made during a 4-min rest period between 8:00 a.m. and 1:00 p.m. Using linear regression
models, adjusted for age, BMI, fasting blood glucose, smoking, cardiac medications, room temperature,
and season. Park et al. (2005b) examined the association between HRV metrics and PM2 5, 03, N02, S02,
CO, BC, and particle number count moving averages in the previous 4, 24, and 48 h. They also estimated
the modifying effects of hypertension, diabetes, ischemic heart disease, and use of hypertensive
medications. Of the pollutants examined, only PM2 5 and 03 were associated with reductions in HRV, and
each pollutant's effect appeared independent of the other. Each 8 |_ig/m3 increase in mean PM2 5
concentration in the previous 48 h was associated with a 20.8% decrease in the component of HRV HF
(95% CI: -34.2% to -4.6%), with larger effects among subjects with hypertension, ischemic heart disease
(IHD), and diabetes. Ozone effects were strongest with the 4 h moving average. The authors state that
since BC concentrations were also associated with adverse changes in HRV, this suggests that traffic
pollution may be particularly toxic (Park et al. (2005b).
In further analyses of the Normative Aging Study cohort, Schwartz et al. (2005b) examined the
hypothesis that adverse changes in HRV due to PM are mediated by an oxidative stress response
(Schwartz et al., 2005a). They examined whether the change in the HF component of HRV associated
with each 10 (ig/m3 increase in 48 h mean PM2 5 was modified by the presence or absence of the allele for
glutathione S-transferase Ml (GSTM1), use of statins, obesity, high neutrophil counts, higher BP, and/or
older age. In subjects without the GSTM1 allele and its protection against oxidative stress, each 10 |_ig/m3
increase in 48 h mean PM2 5 concentration was associated with a 34% decrease in HF (95% CI: -52% to
-9%). There was no association among those with at least one copy of the allele. Obesity and high
neutrophil counts also worsened the effect of PM on HRV regardless of allele (Schwartz et al., 2005a).
Park et al. (2006b) hypothesized that transition metals may be responsible for PM/cardiorespiratory
effects (Park et al., 2006b). Again using the Normative Aging Study cohort, they investigated whether
subjects with two hemochromatosis (HFE) polymorphisms associated with increased iron uptake had a
smaller decrease in HF HRV associated with PM than those subjects without either variant. Each
December 2008
6-12
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
10 (ig/m3 increase in 48 h mean PM25 was associated with a 31.7% decrease in HF (95% CI: -48.1 to
-10.3%) among subjects without either polymorphism, but not among those with the two protective HFE
alleles (Park et al., 2006b).
Again using the Normative Aging Study cohort, Chahine et al. (2007) reported a 10.5% reduction
in SDNN (95% CI: -18.2% to -2.2%) associated with each 10 (ig/m3 increase in the mean 48 h PM2 5
concentration among those without the GSTM1 allele, but only a 2.0% SDNN decrease (95% CI: -11.3%,
8.3%) in those with the allele. This confirmed the PM/HF HRV findings of Schwartz et al. (2005a).
Further, subjects with the long repeat polymorphism in the heme oxygenase-1 (HO-1) promoter had a
greater decline in SDNN associated with each 10 (ig/m3 increase in the mean 48-h PM2 5 concentration
(-8.5%; 95% CI: -14.8% to -1.8%) than those with the short repeat polymorphism in HO-1 (7.4 %
increase; 95% CI: -8.7%, 26.2%). Again, this suggests that PM-HRV changes are mediated, in part, by
oxidative stress (Chahine et al., 2007).
Among those Normative Aging Study subjects with high chronic lead exposure as measured using
X-ray fluorescence of the tibia, each 7 (ig/m3 increase in mean PM2 5 concentration in the previous 48 h
was associated with a 22% decrease in HF HRV (95% CI: -37.4% to -1.7%). Decreases in HF HRV were
also associated with each 2.5 (ig/m3 increase in mean sulfate concentration in the previous 48 h
(22% decrease; 95% CI: -40.4%, 1.6%) and each 16 ppb increase in mean ozone concentration in the
previous 48 h (38% decrease; 95% CI: -54.6% to -14.9%). The authors suggest that these findings are
consistent with an oxidative stress response (Park et al., 2008a). Although this series of studies suggest a
role of oxidative stress in these acute PM/HRV associations, replication by other investigators in other
cities and in other populations will aid interpretations of these findings.
Summary
Omega-3 fatty acid was found to modulate the effect of PM on HRV in a randomized trial
conducted in Mexico City (Romieu et al., 2005). In addition, several analyses of data from the Normative
Aging Study have provided evidence that HRV is modulated by genetic polymorphisms related to
oxidative stress (Chahine et al., 2007; Park et al., 2006b; Schwartz et al., 2005b) or preexisting conditions
such as diabetes, IHD, and hypertension (Park et al., 2005b). Another analysis reports that the HRV-PM
association is more pronounced among those with chronic lead exposure (Park et al., 2008a).
December 2008
6-13
DRAFT-DO NOT QUOTE OR CITE
-------
Table 6-1. Characteristics of epidemiologic/panel studies investigating associations between PM
and changes in HRV.
PM Type,
Exposure
Lag
Study
Subjects
Ambient
Concentration
Mean (SD) **
Recording
Length
SDNN
LF
HF,
r-MSSD
LFHFR
U.S. AND CANADIAN STUDIES
Park et al.
(2005b)
PM2.5, 48-h
avg
497 men
(mean
age = 73[7]
years),
Normative
Aging Study
¦ Boston MA
24-h: 11.4
(8.0) pg/m3
4-min
4
1
1
t
PN, 48-h avg
24-h: 28,942
(13,527)
particles/cm3
1
1
1
BC, 48-h avg
24-h: 0.92
(0.47) pg/m3
4
1
1
t
Liao et al. (2004)
24-h PM10,
lag 1-d
6784(mean
age = 62[6]
years), ARIC
study: MD, NC,
MN
24.3 pg/m3
5-min
1
1
1
Riedeker et al.
(2004b)
In-vehicle
PM2.5 (mass)
9-h avg
9 state
troopers, NC
9-h in-vehicle avg.
23 pg/m3
10-min
1
t
t
Schwartz et al.
(2005b)
BC, 24-h
28 (61-89 y),
12wk
follow-up,
Boston, MA
24-h Median:
1.0 pg/m3
23-min
1
1
t
PM2.5, 24-h
24-h Median:
10 pg/m3
1
1
t
Secondary
PM
(estimated),
24-h
1-h Median:
-1.7 pg/m3
1
1
t
Yeatts et al.
(2007)
24-h PM10-2.5
12 adult
asthmatics,
Chapel Hill, NC
5.3 (2.8) pg/m3
5-min
1
1
1
24-h PM2.5
12.5 (6.0) pg/m3
t
1
t
Wheeler et al.
(2006a)
PM2.5, 4-h
avg
18COPD,
Atlanta, GA
17.8 pg/m3
20-min
t
t
t
t
PM2.5, 4-h
avg
12 Ml, Atlanta,
GA
1
t
1
1
EC, 4 -h avg
18COPD,
Atlanta, GA
2.3 pg/m3
t
Not
presented
Not
presented
Not
presented
EC, 4 -h avg
12 Ml, Atlanta,
GA
1
Not
presented
Not
presented
Not
presented
Dales (2004)
PM2.524-h
avg
(personal)
36 CAD
patients,
Toronto,
Canada
19.9 (13.8) [jg/m3
Not described
Luttmann-Gibson
et al. (2006)
PM2.5, lag 1-d
32 (65+ y),
Steubenville
OH
24-h: 19.7 pg/m3
~30-min.
1
1
1
Sulfate, lag
1-d
24-h: 6.9 pg/m3
1
1
1
Nonsulfate
PM, lag 1-d
24-h: 10.0 pg/m3
1
1
1
EC, lag 1-d
24-h: 1.1 pg/m3
t
1
->
Adaret al.
(2007b)
PM2.5, 24-h
avg
44 (60+y),
diesel bus
riders, St.
Louis, MO
7.7 pg/m3
5-min.
1
1
1
t
BC, 24-h avg
330 ng/m3
1
1
1
t
PC fine
42 particles/cm3
1
1
1
t
PC course
0.02 particles/cm3
t
t
t
1
December 2008
6-14
DRAFT-DO NOT QUOTE OR CITE
-------
PM Type,
Exposure
Lag
Study
Subjects
Ambient
Concentration
Mean (SD) **
Recording
Length
SDNN
up
LF r-MSSD LFHFR
Pope et al.
(2004a)
24-h PM25
(FRM), lag
1-d
88 subjects
(65+ years of
age; 250
p-days),
Utah Valley
23.7 (20.2) pg/m3
24-h
4
1
Sullivan et al.
(2005b)
PM2 5,1,2,
24-h
averages
21 subjects
(65+ years)
with CVD,
Seattle WA
Median:10.7 pg/m3
20-min
13 subjects
(65+ years)
w/out CVD,
Seattle WA
Lipsett (2006)
PM10, lag *
19 CAD
patients (65+
y), 12 wk fu,
Coachella
31.0 and
46.1 pg/m3
5-min F domain;
2-h, 24-h T
domain
4
1 1
PM10-2.5, lag*
None given
1
1 -
PM2.5, lag*
Valley, CA
14 & 23.2 pg/m3
1
1 t
Ebelt et al. (2005)
PM10 — 24 h
16 subjects
with COPD in
Vancouver,
Canada
17 (6) (jg/m3
24-h
1
1
PM10-2.5
(calculated
from PM10
and PM2.5
values)
5.6 (3.0) (jg/m3
t
PM2.5- 24-h
11.4(4.6) (jg/m3
1
1
PM2 5 Sulfate
-24-h
outdoor
2.0 (1.1) (jg/m3
1
-'-
INTERNATIONAL STUDIES
Timonen et al.
(2006)
UF, lags 0-2
days
Stable CHD
patients (65
years of age
and older)
n = 37:
Amsterdam
n = 47: Erfurt
n = 47: Helsinki
Amsterdam: 17,300
particles/cm3
Erfurt: 21,100
particles/cm3
Helsinki: 17,000
particles/cm3
5-min
(Pooled
estimates during
paced breathing
presented to the
right)
1
t 1
AC, lags 0-2
Amsterdam:
1
t 1
days
2100 particles/cm3
Erfurt:
1800 particles/cm3
Helsinki:
1400 particles/cm3
PM2.5, lags
0-2 days
Amsterdam:
20.0 (jg/m3
Erfurt:
23.1 (jg/m3
Helsinki:
12.7 (jg/m3
1
t 1
PM10-2.5, 2-
day lag
Amsterdam:
15.3 (jg/m3
Erfurt:
3.7 (jg/m3
Helsinki:
6.7 (jg/m3
- 1
December 2008
6-15
DRAFT-DO NOT QUOTE OR CITE
-------
PM Type,
Exposure
Lag
Study
Subjects
Ambient
Concentration
Mean (SD) **
Recording
Length
SDNN
LF HF<
r-MSSD
LFHFR
Chan et al.
(2004)
NC0.02-1 1-4 h
9 adults
(19-29) with
lung function
impairment
Taipei, Taiwan
23,407 (19,836)
particles/cm3
5 min
4 4
1
1
10 adults
(42-79) with
lung function
impairment
Taipei, Taiwan
25,529 (20,783)
particles/cm3
1 1
1
1
Chuang et al.
(2005b)
PMi0-0.31-4 h
16CHD
hypertensive
patients, Taipei
37.2 (25.8) (jg/m3
5-min
1 1
1
t
PM2.5-1.01-4 h
12.6 (7.8) (jg/m3
1 1
1
t
PM10-2.51-4 h
Taiwan
14.0 (11.1) |jg/m3
1 1
1
t
PMi0-0.31-4 h
10CHD
patients, Taipei
Taiwan
26.8 (25.9) (jg/m3
1 1
1
->
PM2.5-1.01-4 h
10.9 (8.5) (jg/m3
1 1
1
1
PM10-2.51-4 h
16.4 (10.7) (jg/m3
1 1
1
t
Holguin et al.
(2003)
24-h PM2.5
21 without
hypertension
(60-96 years),
Mexico City
37.2 (13.5) (jg/m3
5-min
1
1
t
13 with
hypertension
(60-88 years),
Mexico City
1
1
t
Romieu et al.
(2005)
24-h PM2.5
(outdoor and
indoor)
50 nursing
home residents
65+ y, Mexico
City
Outdoor: 19.4
(5.7) |jg/m3
Indoor: 18.3
(5.8) (jg/m3
6-min
(Indoor PM2.5,
pre-supplement
phase presented)
1 1
1
Riojas-Rodriguez
et al. (2006)
Personal
PM2.5
30IHD
patients,
Mexico City
Geometric mean:
46.8 (jg/m3
5-min
1
1
Notes: Increases ('), decreases (") and no effects (') in HRV associated with PM concentration are indicated. Statistical significance was not necessary to categorize an effect as an
increase or decrease.
For time domain measures moving average lags up to 24-h were explored. For frequency domain measures lags of 2-h, 4-h and 24-h were explored.
** All concentrations are means, unless otherwise noted.
6.2.1.2. Human Clinical Studies
1 The 2004 PM AQCD cited one study in which parasympathetic stimulation of HRV increased
2 relative to filtered air control following a 2-h exposure with intermittent exercise to fine concentrated
3 ambient particles (CAPs) (average concentration 174 |ig/m3) in both healthy and asthmatic volunteers
4 (Gong et al., 2003a). This effect was observed immediately following the exposure and at 2-days
5 post-exposure, but not at 4-h post-exposure. Although not statistically significant, HRV (total power)
6 increased following exposure to filtered air and decreased following exposure to CAPs. Two new studies
7 have evaluated the effect of PM2 5 CAPs (2-h exposures to concentrations of 20-200 |ig/nr') on HRV in
8 elderly subjects (Devlin et al., 2003; Gong et al., 2004a). In both studies, subjects experienced significant
9 decreases in HRV following exposure to CAPs relative to filtered air exposures. Interestingly, Gong et al.
10 (2004a) found that decreases in HRV were more pronounced in healthy older adults than in those with
December 2008
6-16
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
COPD. In another study, healthy and asthmatic adults were exposed to thoracic coarse CAPs (average
concentration 157 |ig/m3) for 2 h with intermittent exercise (Gong et al., 2004b). HRV was not affected
immediately following the exposure, but decreased in both groups at 4- and 22-h after the end of the
exposure, with greater responses observed in non-asthmatics.
Several additional new human clinical studies have evaluated the effect of PM on HRV in healthy
and health-compromised individuals. Beckett et al. (2005) exposed twelve resting, healthy adults for 2-h
to filtered air and 500 (ig/m3 zinc oxide in the ultrafine (40.4 ± 2.7 nm) and fine (291.2 ± 20.2 nm) modes.
Time and frequency domain parameters of HRV were analyzed immediately following exposure as well
as at 3-, 6-, 11-, and 23-h post-exposure. Neither ultrafine nor fine zinc oxide produced a significant
change in any measure of HRV when compared to filtered air. However, the relevance of zinc oxide to
ambient pollutant particles is unclear.
In a random order crossover human clinical study, Routledge et al. (2006) examined the effects of
ultrafine carbon particles (50 (.ig/nr3) alone and in combination with 200 ppb S02 on HRV among 20
healthy older adults (age 56-75), as well as 20 older adults with coronary artery disease (age 52-74). Five
minute recordings of HRV data were obtained prior to and immediately following the 1-h exposure, as
well as 3 h post-exposure. In healthy subjects, exposure to carbon particles resulted in small increases in
RR-interval, SDNN, rMSSD, and LF power immediately following exposure compared to filtered air
control. At 3 h post-exposure, there were no significant differences in HRV measures between carbon
particle and filtered air exposures. Conversely, S02-induced decreases in HRV were observed at 3 h, but
not immediately following exposure. Concomitant exposure to carbon particles and S02 followed a
pattern similar to that observed with S02 alone, but did not reach statistical significance. Subjects with
coronary artery disease did not experience any significant changes in HRV following exposure to either
pollutant. The authors postulated that this lack of effect may be due to differences in medication between
the two groups, as 70% of subjects with stable angina reported using beta blockers, which are known to
increase cardiac vagal control. In this study, subjects were exposed to pollutants at concentrations similar
to those experienced in urban areas and to which have been associated with mortality and hospital
admissions. The lack of any significant reductions in HRV following exposure to carbon particles is an
important finding, as it provides evidence to suggest that the health effects observed following exposure
to PM may be due to particle constituents other than carbon.
Samet et al. (2007) recently compared the effects of 2-h exposures with intermittent exercise to
ultrafine (average concentration 47 |ig/m3: mean size 0.049 |im). fine (average concentration 120 (ig/m3;
mean size 0.65 |im), and thoracic coarse (average concentration 89 (.ig/nr3; mean size 3.59 |im) CAPs
among healthy subjects between the ages of 18 and 35 years in Chapel Hill, North Carolina. In both the
ultrafine and thoracic coarse studies, a crossover design was used in which each subject was exposed to
December 2008
6-17
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
both PM and filtered air. In the case of the fine PM study, subjects did not serve as their own control, but
were exposed to either PM or filtered air (Ghio et al., 2000). Thoracic coarse fraction CAPs produced a
statistically significant decrease in SDNN 20 h after exposure compared with filtered air exposure. No
statistically significant effects on HRV were observed following exposure to ultrafine PM as measured
during controlled 5-min intervals. However, the authors did observe a significant decrease in SDNN
following exposure to ultrafine PM based on an analysis of the 24-h ambulatory measurements. No
differences were reported in HRV between fine PM exposures and exposures to filtered air. While the
methodologies, exposure criteria, and results have been published separately on the fine PM exposures
included in this investigation (Ghio et al., 2000), only a general summary of the results from the thoracic
coarse and ultrafine exposures are presented in Samet et al. (2007).
In a double-blind, crossover, controlled-exposure study, Peretz et al. (2008b) exposed three healthy
adult volunteers and 13 adults with metabolic syndrome while at rest to filtered air and two levels of DE
(fine PM concentrations of 100 and 200 |ig/m3) in 2-h sessions. HRV parameters were assessed prior to
exposure, as well as at 1-, 3-, 6- and 22-h following the start of exposure, and included both time domain
(SDNN and RMSSD) and frequency domain parameters (HF power, LF power, and the LFHFR ratio).
The authors observed an increase in HF power and a decrease in LFHFR 3 h after the start of exposure to
200 |ig/m3 relative to filtered air. Although these changes were statistically significant (p < 0.05) the
effects were not consistent among the study subjects. No other significant effect of DE on HRV was
observed at either concentration or time point. The authors attributed the lack of consistent effects to the
small and non-homogeneous population and the timing of measurement. There was no difference in either
baseline or diesel-induced changes in HRV parameters between normal individuals and patients with
metabolic syndrome, although the number of normal individuals was quite small (n = 3). It is unclear if
patients with metabolic syndrome were taking any medications. Gong et al. (2008) exposed healthy
(n = 17) and asthmatic (n = 14) adult volunteers to ultrafine CAPs (Los Angeles) for 2 h with intermittent
exercise. Relative to control (filtered air) exposures, UFP exposures (average mass = 100 (ig/m3, average
particle count = 145,000/cm3) among healthy and asthmatic subjects were associated with a transient
decrease in LF power (p < 0.05) 2 h post-exposure.
The results of several new controlled human exposure studies provide limited evidence to suggest
that acute exposure to near ambient levels of PM may be associated with small changes in HRV. Changes
in HRV parameters, however, are inconsistent with some showing increased parasympathetic activity
relative to sympathetic activity and others showing the opposite. Although a direct comparison between
younger and older adults has not been made, PM exposure appears to result in a decrease in HRV more
consistently in healthy older adults (Devlin et al., 2003; Gong et al., 2004a).
December 2008
6-18
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
6.2.1.3. Toxicological Studies
Toxicological studies that examined HR and HRV are presented in the 2004 PM AQCD and overall
demonstrated differing responses, which were collectively characterized as providing limited evidence for
PM-related cardiovascular effects (U.S. EPA, 2004). The studies described that reported HR or HRV
effects following PM exposure were conducted with a variety of particle types (CAPs, diesel, ROFA,
metals), exposure methods (inhalation and IT), and doses (100-3000 |ig/m3 for inhalation; up to 8.3
mg/kg for IT).
SH rats exposed to CAPs in Tuxedo, NY for 4 h (mean PM2.5 concentration 73 |ig/m3: single-day
concentrations 80 and 66 |ig/m3: February and May, 2001, respectively) demonstrated decreased HR
when exposure groups were combined that returned to baseline values when exposure ceased (Nadziejko
et al., 2002). Fine or ultrafine sulfuric acid exposure (mean concentration 225 and 468 |ig/m3.
concentration range 119-299 and 140-750 |ig/m3. respectively) did not induce any HR effects. Another
study demonstrated a trend toward increased HR following a 1- or 4-day PM2 5 CAPs exposure (Wistar
Kyoto rats; 4.5 h/day; Yokohama City, Japan; 5/2004, 11/2004, 9/2005) but the correlation between
change in HR and cumulative PM mass collected over the exposure period was not significant (Ito et al.,
2008).
Decreased SDNN was observed in SH rats exposed via nose-only inhalation to ultrafine CAPs for
two 5-h periods separated by 24 h in the spring (mean mass concentration 202 |ig/m3: 30/10"
particles/cm3; number concentration range 7.12/ 103-8.26z 105 particles/cm3), but not the summer (mean
mass concentration 141 (ig/m3; mean number concentration 2.78/ 10" particles/cm3; number concentration
range 7.76/103—8.87 / 10" particles/cm3) (Chang et al., 2005b). Each of the four animals served as their
own control and mixed effects models were used to determine statistical significance. The estimated mean
PM effects for the SDNN decreases from the start to the end of exposure were 85 to 60% of baseline,
respectively. CAPs effects on rMSSD were less remarkable.
Anselme et al. (2007) used a myocardial infarction model of CHF where the left descending
coronary artery of Wistar rats was occluded to induce ischemia. After 3 months of recovery, rats were
exposed to diesel emissions for 3 h (PM concentration 500 |ig/m3: N02 1.1 ppm; CO 4.3 ppm) and
decreases in rMSSD were observed during the first 2-h of a 3-h diesel exposure, which returned to
baseline values for the last 1-h of exposure. Healthy rats also demonstrated decreased rMSSD when
measured over the entire exposure period. The duration of ventricular premature beat (VPB) attributable
to diesel exposure in CHF rats lasted much longer than the rMSSD change (>5 h post-exposure),
indicating that the HRV response was not driving the increased arrhythmia incidence.
December 2008
6-19
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
A model of premature senescence has been developed by Tankersley et al. (2003), using aged AKR
mice whose body weight abruptly declines ~5 wk prior to death and is accompanied by deficiencies in
other vital physiological function including HR and temperature regulation. When exposed to carbon
black by inhalation (mean concentration 160 |ig/m3: 3 h/day x 3 day), terminal senescent mice responded
with robust cardiovascular effects, including bradycardia and increased HRV indices (rMSSD and SDNN)
(Tankersley et al., 2004). SDNN was also increased in healthy senescent mice exposed to carbon black.
Another measure of HRV, the LFHFR, was similarly elevated in healthy mice with carbon black exposure
compared to the near-terminal mice. These studies indicate that HR regulatory mechanisms are altered in
susceptible mice exposed to PM (sympathetic and parasympathetic changes in healthy senescent mice and
increased parasympathetic influence in terminally senescent mice), which may translate into lowered
homeostatic competence in these animals. Results from the near-terminal group should be interpreted
with caution, as only 3 mice were in this group.
Subsequent research with similar exposure protocols (mean carbon black concentration 159 |ig/m3:
3 h) as the above study used C57BL/6J and C3H/HeJ mice to determine whether an acute PM challenge
can modify HR regulation in two mice strains with differing baseline HR (Tankersley et al., 2007).
Besides the higher HR and lower SDNN and rMSSD in C3H/HeJ compared to C57BL/6J, there were no
carbon black-related changes in HR or HRV. When the C57BL/6J mice (average HR ~80 bpm lower than
C3H/HeJ) were given propanolol (a sympathetic antagonist), carbon black exposure caused an increase in
HR and decrease in rMSSD compared to air during the last 2-h of exposure, indicating withdrawal of
parasympathetic tone. The authors recognize that there may be differences in regional particle deposition
based on strain-specific breathing patterns, which may partially explain the variable HR with carbon
black exposure. Despite this potential shortcoming, this study revealed that inherent autonomic tone,
which is genetically varied between these mouse strains, may affect cardiovascular responses following
PM exposure. In extrapolating these results to humans, individual variation in genetic factors likely plays
some role in PM-induced adjustments in HR control via the ANS.
A recent study in mice (C3H/HeJ, C57BL/6J, and C3H/HeOuJ) examined the effects of a 2-h ozone
(mean concentration 0.584 ppm) pretreatment followed by a 3 h exposure to carbon black (mean
concentration 536 |ig/nr') on HR and HRV measures (Hamade et al., 2008). Data from the both C3H
strains were combined because there were no statistically detectable differences in responses. HR
decreased to the greatest extent during ozone pre-exposure for C3H and C57BL/6J mice that were then
exposed to carbon black. The percent change in SDNN and rMSSD were increased in C3H mice during
ozone pre-exposure and carbon black exposure compared to the filtered air group; however, these HRV
parameters gradually decreased over the duration of the experiment and appeared to be ozone dependent.
Together, these findings led to the conclusion that increases in parasympathetic tone and/or decreases in
December 2008
6-20
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
sympathetic input may explain the observed bradycardia. In a subset of all mice pre-exposed to ozone,
rMSSD remained significantly elevated during the carbon black exposure compared to filtered air. The
results from this study confirm what was observed in Tankersly et al. (2007) in that genetic determinants
affect HR regulation in mice with exposure to air pollutants.
In summary, both increases and decreases in HR have been observed in rats or mice following PM
exposure. Fine or ultrafine sulfuric acid did not result in HR changes in SH rats. Similarly, decreased
SDNN was reported for ultrafine CAPs exposure, but only in the spring and lowered rMSSD was
observed with diesel exposure. In near-terminal senescent mice, HRV responses were robust following
carbon black exposure and represented increased parasympathetic influence. Strain differences in baseline
HR and HRV likely contribute to PM responses. HRV changes with preexposure to ozone and carbon
black appeared to be ozone dependent, although rMSSD remained elevated during PM exposure.
Source Apportionment and PM Components
An additional analysis of CAPs data (Chen and Hwang, 2005; Hwang et al., 2005) was conducted
to link short-term HR and HRV effects to major PM source categories using source apportionment
methodology (Lippmann et al., 2005a). The source categories were: (1) regional secondary sulfate
comprised of high S, Si, and OC (mean 63.41 |ig/m3): (2) resuspended soil characterized by high
concentrations of Ca, Fe, Al, and Si (mean concentration 5.88 |ig/m3): (3) residual oil derived from
power-plant emissions in the Eastern U.S. and containing high levels of V, Ni, and Se (mean
concentration 1.53 (.ig/nr3); and (4) motor vehicle traffic and other unknown sources (34.92 (.ig/nr3)
(Lippmann et al., 2005a). Exposures occurred from 9:00 a.m. to 3:00 p.m., 5 days/wk and the daily time
periods considered in the analysis were 11:00 a.m. to 1:00 p.m., 4:00 to 6:00 p.m., and 1:30 to 4:30 a.m.
PM2 5 mass was associated with a daily interquartile change (difference between third and first quartile of
measured concentrations) of -4.1 beat/min HR during exposure in ApoE" " mice and a similar magnitude
of effect was observed with resuspended soil (-4.5 beat/min). Resuspended soil was also associated with a
HR increase in the afternoon post-exposure (2.6 beat/min); the secondary sulfate factor was linked to
lowered HR in the afternoon post-exposure (-2.5 beat/min). A 6.2% increase in rMSSD collected in the
afternoon post-exposure was associated with the residual oil factor, compared to a 5.6% and 2.4%
decrease in rMSSD at night for secondary sulfate and PM2 5 mass, respectively. Resuspended soil was
associated with a 4.3% increase in rMSSD the night following CAPs exposure. Compared to rMSSD, the
residual oil and secondary sulfate categories showed similar statistically significant parameter estimates
for SDNN.
Recent studies of ECG alterations in mice have indicated a role for PM-associated Ni in driving the
cardiovascular effects. Lippman et al. (2006) presented a posthoc analysis of daily variations in PM2 5 and
December 2008
6-21
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
changes in cardiac dynamics in ApoE" " mice (average exposure: 85.6 |ig/m3: 7/21/ 2004-1/12/2005;
Tuxedo, NY). On the 14 days that the exposed mice had unusually elevated HR, Ni, Cr, and Fe comprised
12.4% of the PM mass, compared to only 1.5% on the other 89 days. Back trajectory analyses indicated
high-altitude winds from the northwest that did not traverse population centers and industrial areas except
the Sudbury Ni smelter in Ontario, Canada. On the 14 days that high HR was observed, the HR elevation
lasted for two days, but only the current day CAPs concentration was statistically significant. SDNN
decreases were statistically significant for all 3 lags (0, 1, 2 days). The GAM regression analysis showed
that only Ni produced a statistically significant effect for HR and SDNN.
6.2.2. Arrhythmia
Epidemiologic and toxicological studies presented in the 2004 PM AQCD provided some evidence
of arrhythmia following exposure to PM. However, a positive association between PM and ventricular
arrhythmias among patients with implantable cardioverter defibrillators was only observed in one study
conducted in Boston, MA, while toxicological studies reported arrhythmogenesis in rodents only
following exposure to ROFA, diesel exhaust, or metals. Recent epidemiologic studies have confirmed the
findings of PM-induced ventricular arrhythmias in Boston, MA, and have also reported increases in
ectopic beats in studies conducted in the Midwest and Pacific Northwest regions of the U.S. In addition,
two studies from Germany have demonstrated positive associations between traffic and combustion
particles and changes in repolarization parameters among patients with ischemic heart disease. Findings
of recent toxicological studies are mixed, with both demonstrated decreases and increases in frequency of
arrhythmia following exposure to CAPs.
6.2.2.1. Epidemiologic Studies
Studies of Arrhythmias Using Implantable Cardioverter Defibrillators
One study reviewed in the 2004 PM AQCD (described below) assessed the effect of short-term
fluctuations in PM on ventricular arrhythmias. Ventricular arrhythmias (i.e. ventricular tachycardia and
ventricular fibrillation), potentially lethal disturbances of the cardiac autonomic nervous system, are
common precursors to sudden cardiac death, and have been examined in several recent studies, described
below.
Previously, Peters et al. (2000) conducted a pilot study in Boston, MA to examine the association
between short-term changes in ambient air pollutant concentrations and increased risk of ventricular
arrhythmias, among a cohort of patients with implantable cardioverter defibrillators (ICD) (Peters et al.,
December 2008
6-22
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
2000). ICDs continuously monitor subject's HR and rhythm and upon detection of an abnormal rhythm
(i.e. rapid HR), they can be programmed to deliver pacing and/or shock therapy to restore normal sinus
rhythm. Those abnormal rhythms that are most severe or rapid are assumed to be due to ventricular
tachycardia or ventricular fibrillation (i.e. life-threatening arrhythmias), and are thus treated with electric
shock. These ICD devices also store information on each abnormal rhythm detected, including the date,
time, and therapy given. Thus, using the date and time of those arrhythmias resulting in electric shock,
Peters et al. (2000) reported an increased risk of ICD shock associated with mean nitrogen dioxide
concentration in the previous two days. Among subjects with frequent events (10 or more during three
years of follow-up) an increased risk of ICD shock was also associated with interquartile range increases
in CO, N02, PM2 5, and BC in the previous 2 days (Peters et al., 2000). Based on these findings, several
studies were conducted to confirm these findings. The study characteristics, as well as the reported effect
estimates and 95% CI associated with each PM metric, are shown in Table 6-2.
Dockery et al. (2005a; 2005b) conducted a follow-up study of n = 203 ICD patients living in
eastern Massachusetts and followed subjects for a longer period of time (up to 7 years) (2005a; 2005b).
They reviewed the ECG for each ICD-detected arrhythmia and included only ventricular arrhythmias
(ventricular fibrillation or ventricular tachycardia). In single pollutant models (generalized estimating
equations), adjusted for season, temperature, relative humidity, day of the week, patient, and a recent prior
arrhythmia, Dockery et al. (2005a; 2005b) reported increased risks of confirmed ventricular arrhythmias
associated with interquartile range increases in every pollutant (PM2 5, BC, sulfate, N02, S02, 03, and
particle number count). None were statistically significant. Among those with a prior ventricular
arrhythmia in the past three days, interquartile range increases in 2 calendar day mean PM2 5, N02, S02,
CO, 03, sulfate, and BC concentrations were all associated with significant and markedly higher risks of
ventricular arrhythmia than among those without a prior arrhythmia. Last, the authors suggested that the
pollutants associated with increased risk of ventricular arrhythmia implicate traffic pollution (Dockery et
al., 2005a; 2005b).
Rich et al. (2005) conducted a case-crossover analysis of these same data to investigate moving
average pollutant concentrations lagged <48 h (Rich et al., 2005). After adjusting for temperature, dew
point temperature, and barometric pressure, they reported significantly increased risk of ventricular
arrhythmia associated with mean PM2 5 and ozone concentrations in the 24 h before the arrhythmia. Each
pollutant effect appeared independent in two pollutant models. In single pollutant models, N02 and S02
also were significantly associated with increased risk, but when included in two pollutant models with
PM2 5, only PM2 5 remained associated with increased risk. They did not, however, find evidence of a
more acute arrhythmic response to pollution (i.e., larger risk estimates associated with moving averages
< 24-h before arrhythmia detection) (Rich et al., 2005).
December 2008
6-23
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Rich et al. (2006b) conducted another case-crossover study in the St. Louis, MO metropolitan area.
Using the same methods as in Boston, they reported increased risk of ventricular arrhythmia associated
mean S02 concentration in the 24-h before the arrhythmia, but not PM2 5 (in single-pollutant models).
Again, they found no evidence of an arrhythmic response with moving average pollutant concentrations
< 24-h before the arrhythmia (Rich et al., 2006a).
In Vancouver, Canada, Vedal et al. (2004) did not find increased risk of ICD shocks associated with
increases in any pollutant concentration (PMi0, Ozone, S02, N02, and CO), after adjusting for temporal
trends, temperature, relative humidity, wind speed, barometric pressure, and proportion of hours with rain.
Secondary analyses among those subjects with two or more discharges per year, and analyses stratified by
season were also null for PMi0, although an association with S02 (lag 2 d) was observed (Vedal et al.,
2004). A case crossover of these dame data examining additional particulate pollutant concentrations
available for a shorter time frome (e.g., PM2 5, sulfate, EC, and OC) also found no increased risk of ICD
shock associated with any pollutant. Rich et al. (2004) did not use the time-stratified control selection
procedure. They used an ambi-directional approach, taking control periods 7 days before and after the
case day when pollution data was available.
The largest ICD study to date examined the risk of ventricular arrhythmias associated with
increases in the daily concentration of numerous particulate and gaseous pollutants in Atlanta, GA
(Metzger et al., 2007). However, they also did not find significant or consistently increased risk of a
ventricular arrhythmia associated with any interquartile range increase in mean daily particulate or
gaseous pollutant concentration at any lag examined (Metzger et al., 2007).
Albert et al. (2007), although not investigating associations with ambient pollution, conducted a
case-crossover study of the association between ventricular arrhythmia and traffic exposure in the hours
before the arrhythmia. They reported an increased risk of ventricular arrhythmia associated with traffic
exposure or driving in the previous hour. They hypothesized that this increased risk was due to either a
stress response from being in a car in heavy traffic, or from traffic-generated air pollution, or a
combination of both (Albert et al., 2007).
Although ICDs detect and treat potentially life-threatening ventricular arrhythmias, other
arrhythmias including episodes of paroxysmal atrial fibrillation (AF) may also be detected. AF is a
common sustained arrhythmia in clinical practice (Go et al., 2001) and a risk factor for stroke
(Prystowsky et al., 1996) and premature mortality (Kannel et al., 1983). In another case-crossover
analysis of data from the Boston ICD study described above, Rich et al. (2006a) identified 91 confirmed
episodes of paroxysmal AF among 29 subjects. In single pollutant models, they reported an increased risk
of AF associated with each 21.7 ppb increase in mean ozone concentration in the hour before the
December 2008
6-24
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
arrhythmia, each 9.4 (ig/m3 increase in PM2 5 in the hour before the arrhythmia, and each 0.83 (ig/m3
increase in BC concentration in the 24 h before the arrhythmia (Table 6-2) (Rich et al., 2006a).
Since 2004, only one study (in Boston), reported adverse associations between PM and ICD
detected ventricular arrhythmias (Berger et al., 2006; Dockery et al., 2005a; 2005b; Dusek et al., 2006;
Ebelt et al., 2005; Peters et al., 2000; Rich et al., 2005; Sarnat et al., 2006c), while other studies done
elsewhere did not (Dusek et al., 2006; Metzger et al., 2007; Rich et al., 2004; Vedal et al., 2004). A wide
range in exposure lags has been reported in the Boston study (3 h to 3 days) (Dockery et al., 2005a;
Dockery et al., 2005b; Rich et al., 2005). It is not clear why these findings are inconsistent. Rich et al.
(2005) reported that use of the mean pollutant concentration from the specific 24 h before the arrhythmia
rather than just the day of the arrhythmia, resulted in less exposure misclassification and less bias towards
the null, possibly explaining the lack of association when using just the day of ICD discharge and daily
PM concentrations (Rich et al., 2005). Other reasons for the inconsistent findings may include differing
degrees of exposure misclassification within each study or city due to differences in PM composition and
pollutant mixes (e.g., less transition metals and sulfates in the Pacific Northwest than the Northeast U.S.),
and differences in the size of study areas (Boston: within 40 km of PM2 5 monitoring site; Vancouver:
Lower Mainland of British Columbia 90 km east of Vancouver). Studies of ventricular arrhythmia and
PM concentration in patients with ICDs are summarized in Table 6-2.
Table 6-2. Studies of ventricular arrhythmia and ambient PM concentration, in patients with
implantable cardioverter defibrillators.
Reference
Outcome and
Sample Size
Study Design
and Analytic Copollutants
Method
PM
Metric
Ambient
Concentration
Lag and its
Increment
Units
OR
95%
Confidence
Interval
PM2.5
Daily Median:
2 day
1.08 0.96,1.22
10.3 (jg/m3
6.9 (jg/m3
BC
Daily Median:
2 day
1.11 0.95,1.28
0.98 (jg/m3
0.74 (jg/m3
Sulfate
Daily Median:
2 day
1.05 0.92,1.20
2.55 (jg/m3
2.04 (jg/m3
Dockery
(2005a;
2005b)
Eastern MA
N = 670 days
with a 1 EGM
confirmed
ventricular
arrhythmias
among n = 84
subjects
Generalized
estimating
equations
Lags Evaluated:
2 calendar day
means
03
Particle
Number
Daily Median:
29,300
particles/cm3
2 day 1.14 0.87,1.50
19,120
particles/cm3
Rich et al.
(2005)
Eastern MA
N = 798 EGM
confirmed
ventricular
arrhythmias
among n = 84
subjects
Time-stratified
case-crossover
study.
Conditional
logistic regression
Lags Evaluated:
3, 6, 24, 48 h
moving averages
N02,CO, S02, PM25
03
Daily Median:
9.8 (jg/m3
24-h moving 1.19 1.02,1.38
average
7.8 (jg/m3
BC
Daily Median:
0.94 (jg/m3
24-h moving 0.93 0.74,1.18
average
0.83 (jg/m3
December 2008
6-25
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Outcome and
Sample Size
Study Design
and Analytic
Method
Copollutants
PM
Metric
Ambient
Concentration
Lag and its
Increment
Units
OR
95%
Confidence
Interval
Rich et al.
(2006b)
St. Louis
N = 139 EGM
confirmed
ventricular
arrhythmias
among n = 56
subjects
Time-stratified
case-crossover
study.
N02, CO, so2,
03
PM2.5
Daily Median:
16.2 (jg/m3
24-h moving
average
9.7 (jg/m3
0.95
0.72,1.27
metro area
Conditional
logistic regression
Lags Evaluated:
EC
Daily Median:
0.6 (jg/m3
24-h moving
average
0.5 (jg/m3
1.18
0.93,1.50
6,12,24, 48 h
moving averages
Organic
Carbon
Daily Median:
4.0 (jg/m3
24-h moving
average
2.3 (jg/m3
1.08
0.81,1.43
Vedal et al.
(2004)
Vancouver,
BC.CA
N = 257 days
with a 1 ICD
shock among
n = 50 subjects
Generalized
estimating
equations
Lags Evaluated:
0,1, 2, 3 daily
moving average
no2, CO, so2,
03
PM10
Daily Median:
11.6 (jg/m3
Lag Day 0
5.6 (jg/m3
1.00*
0.82,1.19*
Rich et al.
(2004)
N = 77 to 98
days with with a
1 ICD shock
among n = 34
subjects
Ambi-directional
case-crossover
study.
Conditional
logistic regression
Lags Evaluated:
0,1, 2, and 3 day
moving averages
no2, CO, so2,
03
PM2.5
Daily Mean:
8.2 (jg/m3
Lag Day 0
5.2 (jg/m3
1.0t
0.9,1.1|
Vancouver,
BC.CA
PM10
Daily Mean:
13.3 (jg/m3
Lag Day 0
7.4 (jg/m3
0.9|
0.5,1.5f
EC
Daily Mean:
0.8 (jg/m3
Lag Day 0
0.4 (jg/m3
1.1T
0.9,1.3f
Organic
Carbon
Daily Mean:
4.5 (jg/m3
Lag Day 0
2.2 (jg/m3
1.1T
0.9,1.3f
Sulfate
Daily Mean:
1.3 (jg/m3
Lag Day 0
0.9 (jg/m3
0.9|
0.7,1.2f
Metzger et
al. (2007)
Atlanta, GA
N = 6287 EGM
confirmed
ventricular
arrhythmias
among n = 518
subjects
Generalized
estimating
equations
Lags Evaluated:
0,1, and 2 day
moving averages
no2, CO, so2,
03
PM2.5
Daily Median:
16.2 (jg/m3
24-h moving
average
10 (jg/m3
1.00
0.95,1.04
PM10
Daily Median:
26.4 (jg/m3
24-h moving
average
10 (jg/m3
1.00
0.97,1.03
PM10-2.5
Daily Median:
8.7 (jg/m3
24-h moving
average
5 (jg/m3
1.03
1.00,1.07
PM2.5 EC
Daily Median:
1.4 (jg/m3
24-h moving
average
1 (jg/m3
1.01
0.98,105
PM2.5
Organic
Carbon
Daily Median:
3.9 (jg/m3
24-h moving
average
2 (jg/m3
1.01
0.98,1.03
PM2.5
Sulfate
Daily Median:
4.1 (jg/m3
24-h moving
average
5 (jg/m3
0.99
0.93,1.06
PM2.5
water
soluble
elements
Daily Median:
0.022 (jg/m3
24-h moving
average
0.03 (jg/m3
0.95
0.90,1.00
Estimated from Figure 3 (Vedal et al., 2004)
t Estimated from Figure 3 (Rich et al., 2004)
Ectopy Studies Using ECG Measurements
A few panel studies have used ECG recordings to evaluate associations between ectopic beats
(ventricular or supraventricular) and mean particulate concentrations in the previous hours and/or days
December 2008
6-26
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
(Berger et al., 2006; Ebelt et al., 2005; Sarnat et al., 2006c). Ectopic beats are defined as extra cardiac
depolarizations and are the most common disturbance in heart rhythm. Ectopic beats are usually benign,
and may present with or without symptoms, such as palpitations or dizziness. When three or more occur
in succession, this is called a non-sustained ventricular tachycardia. Sustained ventricular tachycardias are
the arrhythmias investigated in the ICD studies described above. Sarnat et al. (2006c) conducted a panel
study among 32 nonsmoking older adults residing in Steubenville, OH (Sarnat et al., 2006c). In this study,
the median daily PM2.5 concentration was 17.7 (ig/m3. The median daily sulfate concentration was
5.7 |ag/m3. and the median daily EC concentration was 1.0 (ig/m3. They used logistic regression models to
examine lagged effects of 1-10 day moving average concentrations of PM25, sulfate, EC, 03, N02, and
S02. In single pollutant models, each 10.0 (ig/m3 increase in 5-day mean PM2.5 concentration was
associated with increased risk of supraventricular ectopy (OR = 1.42; 95% CI: 0.99, 2.04), but not
ventricular ectopy (OR = 1.02; 95% CI: 0.63-1.65). Similarly, increased risk of supraventricular ectopy,
but not ventricular ectopy, was associated with each interquartile range increase in 5-day mean sulfate and
ozone concentration (Sarnat et al., 2006c).
Ebelt et al. (2005) conducted a repeated measures panel study of 16 patients with COPD in the
summer of 1998 in Vancouver, British Columbia (Ebelt et al., 2005). Their goal was to evaluate the
relative impact of ambient and non-ambient exposures to PM2 5, PM10, and PM10.2.5 on several health
measures. The mean PM2 5 concentration during this study was 11.4 (ig/m3. Using mixed models with
random subjects effects to investigate only same day PM concentrations, Ebelt et al. (2005) reported an
increase in supraventricular ectopic beats associated with same day ambient exposures to each PM size
fraction (Ebelt et al., 2005); however, results were presented in figures only.
Berger and colleagues (2006) conducted a panel study of 57 men with coronary heart disease living
in Erfurt, Germany (Berger et al., 2006). Using 24-h ECG measurements made once every 4 weeks, they
studied associations between runs of supraventricular and ventricular tachycardia and lagged
concentrations of PM2 5, UFP (0.01-0.1 (im), ACP (0.1-1.0 (im), S02, N02, CO, and NO. Using
generalized additive models, as well as poisson and linear regression models, they reported increases in
supraventricular tachycardia and the number of runs of ventricular tachycardia associated with 5-day
mean PM2 5, UFP counts (0.01-0.1 andACP counts (0.1-1.0 |_im), after controlling for long-term time
trend, weekday, air temperature, relative humidity, and barometric pressure. They found these associations
at all lags evaluated (during ECG recording, 0-23 h before, 24-47 h before, 48-71-h before, 72-95 h
before, and 5 day mean), but the largest effect estimates were generally associated with the 24-47 h mean
and the 5 day mean (Berger et al., 2006).
These studies of ectopic beats and runs of supraventricular and ventricular tachycardia, captured
using ECG measurements, all report positive associations. They are consistent with the Boston ICD
December 2008
6-27
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
studies described above, although they report findings in other regions (Midwest U.S., Pacific Northwest,
and Erfurt Germany). A range of lags were investigated (0-10 days) with the strongest effects observed
for either the 5-day mean or same day PM concentrations. Taken together, these ICD studies and ectopy
studies provide some evidence of an arrhythmic response to PM, although further study is needed to
understand the discrepancy in ICD study findings.
ECG Abnormalities indicating Arrhythmia
No investigations of the relationship of PM concentration and ECG abnormalities indicating
arrhythmia were conducted prior to 2002 and thus were not included in the 2004 PM AQCD.
Abnormalities in the myocardial substrate, myocardial vulnerability, and resulting repolarization
abnormalities are believed to be key factors contributing to the development of arrhythmogenic
conditions such as those discussed above. These abnormalities include ECG measures of repolarization
such as QT duration (time for depolarization and repolarization of the ventricles), T-wave complexity (a
measure of repolarization morphology), and T-wave amplitude (height of the T-wave). Abnormalities in
repolarization may also identify subjects potentially at risk of more serious events such as sudden cardiac
death (Atiga et al., 1998; Berger et al., 1997; Chevalier et al., 2003; Okin et al., 2000; Zabel et al., 1998).
Recent studies of changes in these measures following acute increases in air pollution are described
below.
Two studies conducted in Erfurt, Germany (Henneberger, 2005; Yue et al., 2007) examined the
association between measures of repolarization (QT duration, T-wave complexity, T-wave amplitude,
T-wave amplitude variability) and particulate air pollution. Henneberger et al. (2005) conducted a panel
study of 56 males with ischemic heart disease (Henneberger, 2005). Each subject was measured every 2
weeks for 6 months. During the study, the median daily PM2 5 concentration was 14.9 (ig/m3. The median
EC concentration was 1.8 (ig/m3, while the median OC concentration was 1.4 |_ig/nr\ The median count of
UFP counts (0.01-0.1 |_im) was 11,444 particles/cm3, while the median count of ACP (0.1-1.0 |_im) was
1238 particles/cm3. Using generalized additive models adjusted for subject, long-term time trend,
temperature, relative humidity, barometric pressure, and weekday, they examined the change in these
ECG parameters associated with the mean pollutant (UFP, ACP, PM2 5, OC, and EC) concentrations 0-5,
6-11, 12-17, 18-23, and 0-23 h before, and 2-5 days before the ECG measurement. Significant decreases
in T-wave amplitude were associated with PM2 5 mass, UFP, and ACP. Each 16.4 (ig/m3 increase in the
mean PM2 5 concentration in the previous 5 h was associated with a 6.46 |_iV decrease in T-wave
amplitude (95% CI: -10.88 to -2.04). Each 0.7 (ig/m3 increase in the mean OC concentration in the
previous 5 h was associated with a 4.15 ms increase in QT interval (95% CI: 0.22, 8.09). There was a
December 2008
6-28
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
similar sized effect for 24-h mean OC concentration. Significant increases in the variability of T-wave
complexity were also associated with acute increases in EC and OC concentration (Henneberger, 2005).
Yue et al. (2007) then used positive matrix factorization to identify 5 sources of ambient PM
(airborne soil, local traffic-related UFP, combustion generated aerosols, diesel traffic-related particles, and
secondary aerosols). Using similar statistical models, they examined the association between these same
repolarization changes and incremental increases in the mean concentration of each particle source in the
24-h before the ECG measurement. They also examined associations with CRP and vWF concentrations
in the blood. Both UF from local traffic and diesel particles from traffic had the strongest associations
with repolarization parameters (Yue et al., 2007).
These two analyses demonstrate associations between PM pollution and repolarization changes, at
lags of 5 h to 2 days. Moreover, the findings from the Yue et al. (2007) study demonstrate a potential role
of traffic particles/pollution.
6.2.2.2. Toxicological Studies
The ECG of animal research models frequently exhibit different characteristics than that of
humans. Mice and rats are notable in this regard, as they do not have an isoelectric ST-segment typical of
larger species, likely owing to their rapid heart rates (-600 and -350 bpm, respectively). However, the
ultimate function of the pumping heart is conserved and reflected by the ECG in a remarkably consistent
manner across species. Thus, atrial depolarization causes an electrical inflection represented by the
P-wave, ventricular depolarization elicits the QRS complex, and the biopotential recovery of the
ventricles is reflected by the T-wave.
The earliest indication that there may be cardiovascular system effects of PM came from ECG
studies in susceptible animal models (rats with pulmonary hyptertension (Watkinson et al., 1998) and
dogs with coronary occlusion (Godleski et al., 2000), which were summarized in the 2004 PM AQCD.
However, a study of dogs exposed to ROFA did not demonstrate ECG changes, perhaps due to differences
in disease state, as these were the oldest dogs in the colony with clinical signs of preexisting, naturally
occurring heart disease (Muggenburg et al., 2000). Much of the research conducted since the release of
the last PM AQCD has been focused on exploring susceptibility or varying exposure methodologies, with
little new evidence into the mechanisms for ECG changes of inhaled PM.
Wellenius et al. (2004) used a susceptible model that was previously shown to produce significant
results with exposures to ROFA (Wellenius et al., 2002) to examine ECG-related PM effects. Using an
anesthetized model of post-infarction myocardium sensitivity, Wellenius and colleagues tested the effects
of concentrated PM2 5 on the induction of spontaneous arrhythmias in rats (Sprague Dawley; 1-h
exposure; Boston, MA; July 2000 to January 2003; mean mass concentration 523.11 |ig/nr': range of mass
December 2008
6-29
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
concentration 60.3-2202 (ig/m3). CAPs caused a statistically significant decrease (67.1%) in ventricular
premature beat (VPB) frequency during the post-exposure period in rats with a high number of
pre-exposure VPB. No interaction was observed with co-exposure to carbon monoxide (35 ppm), which
reduced VPB frequency during the exposure period when administered alone. When further analyses were
conducted to determine whether the CAPs number concentration or the mass concentration of any single
element was a predictor of VPB frequency, no significant results were found. In a follow-up publication,
results from the analysis of supraventricular ectopic beats (SVEB) were provided (Wellenius et al.,
2006b). A decrease in the number of SVEB was observed with CAPs (mean mass concentration
645.7 (.ig/nr3; range of mass concentration 78.0-2202.5 |ig/m3) or CO (average concentration 37.9 ppm)
compared to filtered air. Furthermore, an increase in CAPs number concentration of 1000 particles/cm3
was associated with a 3.3% decrease in SVEB frequency. The findings of decreased ventricular
arrhythmia differ from those observed following ROFA exposure in the same animal model in that an
increased frequency of premature ventricular complexes was observed with ROFA, albeit the exposure
concentration was >3000 |ig/m3 (Wellenius et al., 2002). It is difficult to directly compare the result of
these studies due to differences in exposure concentrations and the nature of CAPs exposures
(i.e., varying daily PM composition), but collectively they may suggest an important role for the soluble
components of PM, including transition metals, as only ROFA induced increases in ventricular arrhythmia
occurrence.
Anselme and colleagues (2007) exposed rats with and without induced CHF to diesel emissions for
3 h (PM concentration 500 |ig/m3: N02 1.1 ppm; CO 4.3 ppm). While no dramatic change was noted in
HR, prominent increases in the incidence of VPB were observed in CHF rats, which lasted at least 4-5 h
after the exposure ceased. It is interesting to contrast the work of Anselme with the studies by Wellenius
et al. (2002; 2004; 2006a), as the arrhythmia incidence in the acute infarction model was greatest with
ROFA, while the CHF model demonstrated sensitivity to diesel exposure. However, several differences in
the research designs preclude strong comparisons.
In older rats (Fisher 344; -18 mo) exposed to PM2 5 CAPs in Tuxedo, NY (4 h; mean concentration
180 |ig/m3: single-day concentrations 161 and 200 (.ig/nr3; August 2000), the frequency of delayed beats
was greater than in rats exposed to air (Nadziejko et al., 2004). The majority of these beats were
characterized as pauses (a delay of 2.5 times the adjacent interbeat intervals) rather than premature beats.
When the same animals were exposed to generated ultrafine carbon (mean concentration 890 |ig/m3:
single-day concentrations 500 and 1280 (.ig/nr3; 30-50 nm MMAD) or S02 (1.2 ppm), no significant
differences were observed in arrhythmia frequency between air controls and exposed animals. The
authors also report using the same protocol for young Wistar rats (concentration 215 |ig/m3) and very few
arrhythmias were observed, thus precluding statistical analysis. The results of this study indicate 1)
December 2008
6-30
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
involvement of the sino-atrial node, as the observed arrhythmias were mostly of a delayed nature and 2)
particle size and PM2 5 constituents may play a role in these effects.
Using ApoE"" mice on a high-fat diet as a model of pre-existing coronary insufficiency (Caligiuri et
al., 1999), Campen and colleagues studied the impact of inhaled diesel and gasoline emissions and road
dust (6 h/day x 3 day) on ECG morphology (2005; Campen et al., 2006). Moreover, the investigators used
a high efficiency particle filter to compare the whole exhaust with an atmosphere containing only the
gaseous components. For gasoline emissions, the PM-containing atmosphere (PM mean concentration
61 |ig/m3: NOx mean concentration 18.8 ppm; CO mean concentration 80 ppm) induced T-wave
morphological alterations, while the PM-filtered atmosphere did not (Campen et al., 2006). Moreover,
resuspended road dust (PM25), at up to 3500 (.ig/nr1 had no impact on ECG. For diesel emissions at higher
concentrations (PM mean concentration 512, 770, or 3634 |ig/m3: NOx mean concentration 19, 105, 102
ppm for low whole exhaust, high PM filtered, and high whole exhaust, respectively; CO levels not
provided), dramatic bradycardia, decreased T-wave area, and arrhythmia (atrioventricular-node block and
VPB) were only observed in ApoE" " mice exposed to the high filtered and high whole exhaust groups
(Campen et al., 2005). These effects remained after filtration of PM, suggesting that the gaseous
components of the whole DE drove the cardiovascular findings. These results contrast, in that the gasoline
emissions required particles to induce T-wave changes, whereas the PM-filtered diesel emissions
produced altered ECG responses. However, the diesel emissions contained much greater PM
concentrations compared to the gasoline emissions.
The above studies demonstrate mixed results for arrhythmias. Wellenius et al. (2004; 2006b)
showed decreased frequency of VPB and SVEB following CAPs exposure in rats with induced MI (>12-h
prior to exposure). In contrast, rats with a MI model of chronic heart failure (3-month recovery) had
increased incidence of VPB with diesel exposure (Anselme et al., 2007). One study reported increased
frequency of premature beats in older rats exposed to CAPs, which were not observed with ultrafine
carbon particles (Nadziejko et al., 2004). As for ECG morphology changes, T-wave alterations were
reported for gasoline emissions that were absent when the PM was filtered (Campen et al., 2006).
However, for DE, increased atrioventricular-node block, VPB, and decreased T-wave area were observed
with whole exhaust and remained after filtration of PM, indicating that the gases were responsible for the
effects (Campen et al., 2005).
6.2.3. Ischemia
Although no evidence from epidemiologic or human clinical studies of PM-induced myocardial
ischemia was included in the 2004 PM AQCD, one toxicological study was cited that observed ST-
December 2008
6-31
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
segment changes in dogs following a 3 day exposure to CAPs. In epidemiologic studies published since
the 2004 PM AQCD, associations have been demonstrated between PM and ST-segment depression, and
one new human clinical study reported significant increases in exercise-induced ST-segment depression
among men with prior MI following a controlled exposure to diesel exhaust. Results from recent
toxicological studies confirm the findings presented in the 2004 PM AQCD and provide coherence and
biological plausibility for the effects observed in epidemiologic and human clinical studies.
6.2.3.1. Epidemiologic Studies
ECG Abnormalities Indicating Ischemia
There were no studies of ECG abnormalities indicating ischemia reviewed in the 2004 PM AQCD.
The ST-segment duration is typically in the range of 0.08 to 0.12 sec (80 to 120 ms). A depression in the
ST-segment may indicate coronary ischemia, while an elevation may indicate MI. Several short-term
exposure studies of air pollution investigated the association of ST-segment depression with PM
concentration.
Gold et al. (2005) studied 24 elderly residents of Boston, MA (aged 61-88 years) residing at or near
an apartment complex that was ~ 0.5 km from an air pollution monitoring station. A protocol of
continuous Holter monitoring including 5 mins of rest, 5 mins of standing, 5 mins of outdoor exercise, 5
mins of rest, and then 20 cycles of paced breathing was done up to 12 times for each subject (n = 269
ECG measurements for analysis). From these ECG measurements, they identified occurrences of ST-
segment depression and examined whether mean BC, CO, and PM2 5 concentrations in the previous 5 and
12-h were associated with ST-segment depression. The median 5 h mean BC concentration was
1.28 |_ig/m3and the median 12-h mean BC concentration was 1.14 (ig/m3. The median 5 h mean PM2 5
concentration was 9.5 |_ig/m3. and the median 12 hour mean PM2 5 concentration was 9.8 |_ig/m3. Using
single pollutant, conditional linear mixed regression models, including a cubic term for current hourly
temperature and a linear trend of time, Gold et al. (2005) reported that the mean BC concentrations in the
5 and 12 h before testing predicted ST-segment depression in most portions of the protocol. However,
these effects were strongest in the post-exercise periods. For example, during the post-exercise rest
period, each 10th-90th percentile increase (1.59 (.ig/ni3) in the mean 5 hour BC concentration was
associated with a-0.11 mm ST-segment depression (95% CI: -0.18 to -0.05). In two pollutant models, CO
did not appear to confound this association. These findings suggest traffic-generated particulate pollution
may be associated with ST-segment depression (Gold et al., 2005).
Previously, Pekkanen et al. (2002) conducted a panel study of 45 subjects with stable coronary
heart disease living in Helsinki, Finland. Each subject had biweekly sub-maximal exercise testing for 6
December 2008
6-32
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
months (n = 342 exercise tests with 72 exercise-induced ST-segment depressions). During this study, the
median daily PM2 5 concentration was 10.6 (ig/m3, and the median daily count of ACP (ACP: 0.1-1.0 |_im)
was 1200 particles/cm3. Using logistic regression and generalized additive models adjusting for time
trend, temperature, relative humidity, and HR change during testing, they examined the risk of ST-
segment depression associated with mean pollutant concentrations (UFP, ACP, PMi, PM2 5, PM2 5.10, N02,
CO) in the previous 24-h, and the 3 previous lagged 24-h periods. Each 7.9 (ig/m3 increase in mean PM2 5
concentration, lagged 2 days, was associated with significantly increased risk of ST-segment depression
>0.1 mV (OR: 2.84 [95% CI: 1.42-5.66]). Each 760 particles/cm3 increase in the count of ACP, lagged 2
days, was also associated with significantly increased risk of ST-segment depression >0.1 mV (OR: 3.29
[95% CI: 1.57-6.92]). Similarly sized increased risks of ST-segment depression were also found for other
particulate pollutants, including PMi0.2.5, PMi, and the counts of UFP (0.01-0.1
This same research group, then conducted a principal components analysis to identify five PM2 5
sources (crustal, long range transport, oil combustion, salt, and local traffic) (Lanki et al., 2006b). Using
similar statistical models, each 1 (ig/m3 increase in "local traffic" particle concentration, lagged 2 days,
was associated with increased risk of ST-segment depression (OR: 1.53 [95% CI: 1.19-1.97]). Similarly,
each 1 (ig/m3 increase in "long range transport" particle concentration was also associated with increased
risk of ST-segment depression (OR: 1.11 [95% CI: 1.02-1.20]). No significant associations for other
sources were reported for any lag time. These studies demonstrate associations between PM pollution and
ST-segment depression at lags of 5 hours to 2 days. Morever, these findings demonstrate a potential role
for traffic (Gold et al., 2005) and long-range transported PM (Lanki et al., 2006a).
6.2.3.2. Human Clinical Studies
Among a group of 20 men with prior MI, Mills et al. (2007) found that DE (300 (.ig/m3 PM)
significantly increased exercise-induced ischemic burden during exposure, calculated as the product of
exercise duration and change in ST-segment amplitude. The mechanism by which DE induced the
exacerbation of ischemic burden remains unclear, and appears to be unrelated to impaired vasodilation.
However, the authors suggest that this discrepancy may be due to the timing of the vascular assessment,
as measures of blood-flow were taken 5 hours after the observed increase in ischemic burden.
6.2.3.3. Toxicological Studies
There were no toxicological studies cited in the 2004 PM AQCD that directly examined myocardial
ischemia. One study was reviewed in the 2004 PM AQCD that evaluated ST-segment changes in dogs
December 2008
6-33
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
during occlusion and following a 3-day exposure to Boston CAPs reported increased magnitude and
decreased time to ST-segment elevation (Godleski et al., 2000).
In the first study of its kind, Cozzi et al. (2006) exposed ICR mice to ultrafine PM (100 (ig via
intratracheal instillation), followed by ischemia/reperfiision injury to the left anterior coronary artery.
Both the area-at-risk (the region of tissue perfused by the left anterior descending coronary artery) and the
infarct size were measured 2 h following reperfiision and while the area-at-risk was not affected by PM
exposure, the infarct size was nearly doubled in mice who received ultrafine PM. Increases in infarct size
were associated with increased myocardial neutrophil density in the infarct zone and lipid peroxidation in
the myocardium.
A more recent study in dogs (female mixed-breed canines, 14-18 kg) evaluated myocardial blood
flow during myocardial ischemia following 5-h PM2 5 CAPs exposures in Boston (daily mass
concentration range 94.1-1556.8 (ig/m3; particle number concentration range 3000-69300 particles/cm3;
BC concentration range 1.3-32.0 (.ig/nr3) (Bartoli et al., 2008). Similar methods were used for the
coronary occlusion and exposure method as those described in Wellenius et al. (2003). In addition,
silicone catheters were chronically implanted in the left atrium and descending aorta for fluorescent
microsphere injection and blood flow measurements, respectively. Using a crossover design, 4 animals
were exposed to alternating CAPs and filtered air two times (8 total CAPs exposures). Immediately after
exposure, dogs underwent two 5-min occlusions of the left anterior descending coronary artery with
injection of microspheres (15 |im diameter) after 3 min of ischemia during the second occlusion. Blood
was sampled from the descending aorta to determine the absolute flow to tissue samples. Post-mortem
analysis of cardiac tissue and blood samples allowed for quantification of microspheres using flow
cytometry. CAPs-exposed dogs had decreased total myocardial blood flow and increased coronary
vascular resistance during coronary artery occlusion that was greatest in tissue within or near the ischemic
zone. The rate-pressure product (product of HR and systolic BP) was unchanged in animals exposed to
CAPs during occlusion, indicating that cardiac metabolic demand was not altered. The multi-level linear
mixed models demonstrated that myocardial blood flow and coronary vascular resistance during
occlusion were inversely and significantly associated with CAPs mass concentration, particle number
concentration, and BC concentration, with the strongest effects observed with particle number
concentration. The results of this study provide evidence that exacerbation of myocardial ischemia
following PM exposure is due to reduced myocardial blood flow.
The two studies described above provide evidence that PM can induce greater myocardial
responses following ischemic events, as demonstrated by increased infarct size, decreased myocardial
blood flow and increased coronary vascular resistance. The results indicate that collateral vessels may be
dysfunctional following PM exposure.
December 2008
6-34
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
A study that examined ECG changes in dogs (female; retired mongrel breeder dogs) following
CAPs exposure (mean mass concentration 345 |ig/m3: mass concentration range 161-957 |ig/m3:
September 2000 to March 2001; Boston, MA) and left anterior descending coronary artery occlusion as
an indicator of myocardial ischemia reported changes in ST-segment (Wellenius et al., 2003). The
experimental protocol was a 6-h exposure to CAPs or filtered air via tracheostomy, followed by a
preconditioning occlusion (5 min), rest interval (20 min), and the experimental occlusion (5 min).
Increased ST-segment elevation (estimated as the area under the response curve) was observed following
PM2 5 during the experimental occlusion period compared to filtered air. Furthermore, peak ST-segment
elevation attributable to CAPs was reported with the experimental occlusion, which remained elevated
24-h post-exposure. Ventricular arrhythmias were rarely observed during occlusion and when observed,
were unrelated to CAPs exposure. The results from this study support those observed previously
(Godleski et al., 2000) and provides greater support that enhanced myocardial ischemia occurs relatively
quickly following PM exposure (within hours).
PM Components
The Wellenius et al. (2003) study employing dogs also attempted to link ST-segment changes with
four CAPs elements (Si, Ni, S, and BC) as tracers of PM2 5 sources in Boston. In the multivariate
regression analyses, peak ST-segment elevation and integrated ST-segment change were significantly
associated with only the mass concentration of Si (Si mean concentration 8.17 (.ig/nr3; Si concentration
range 2.31-13.93 |ig/m3). In the univariate regression analyses, Pb also demonstrated a significant
association for both ST-segment measures, although the p-value was greater than that observed with Si.
6.2.4. Vasomotor Function
Perhaps the most noteworthy new health-related revelation in the past six years with regards to PM
exposure is that the systemic vasculature may be a target organ. The vasculature of all tissues is lined with
endothelial cells that will naturally encounter any systemically absorbed toxin. The endothelium
(1) maintains barrier integrity to ensure fluid compartmentalization, (2) communicates dilatory and
constrictive stimuli to vascular smooth muscle cells, and (3) recruits inflammatory cells to injured
regions. Smooth muscle cells lie within the layer of endothelium and are crucial to the regulation of blood
flow and pressure. In states of injury and disease, both cell types can exhibit dysfunction and even
pathological responses.
Endothelial dysfunction is a factor in many diseases and may contribute to the origin and/or
exacerbation of perfiision-limited diseases, such as MI or IHD, as well as hypertension. Endothelial
December 2008
6-35
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
dysfunction is also a characteristic feature of early and advanced atherosclerosis. A primary outcome of
endothelial dysfunction is impaired vasodilatation, frequently due to uncoupling of NOS. It is this
uncoupling that appears central to impaired vasodilation and thus endothelial dysfunction.
One human clinical study cited in the 2004 PM AQCD reported a decrease in bronchial artery
diameter (BAD) among healthy adults following exposure to CAPs in combination with ozone.
Conclusions based on this finding were limited due to the concomitant exposure to ozone as well as a lack
of published results from epidemiologic and toxicological studies. Recent human clinical studies have
provided support to the findings described in the 2004 PM AQCD, with changes in vasomotor function
observed following controlled exposures to diesel exhaust, elemental carbon particles, and indoor air
particles. In addition, epidemiologic studies have observed associations between PM and decreases in
BAD and flow mediated dilatation in healthy adults and diabetics. These findings are further supported by
a large body of new toxicological evidence of impaired vasodilation following exposure to PM.
6.2.4.1. Epidemiologic Studies
O'Neill et al. (2005a) examined the association between 2 measures of vascular reactivity
(nitroglycerin mediated reactivity and flow-mediated reactivity) and ambient mean particulate pollutant
concentration (PM2 5, sulfates, BC, particle number count [PNC]) on the same and previous few days
(O'Neill et al., 2005a). They studied a panel of 270 subjects with diabetes or at risk for diabetes, who
lived in the greater Boston metropolitan area. The mean PM2.5 concentration during this study was
11.5 (ig/m3. Using linear regression models adjusted individually for age, gender, body mass index, and
race, O'Neill et al. (2005a) estimated the change in vascular reactivity associated with moving average
pollutant concentrations across the same and previous 5 days. Interquartile range (values not reported)
increases in the mean PM2 5 concentration, BC concentration, and PNC over the previous 6 days were
associated with decreased vascular reactivity among diabetics, but not among subjects at risk for diabetes.
For sulfates, the mean concentration on lag day 0, lag day 1, and the 3-day, 4-day, and 5-day moving
averages all were associated with similarly sized reductions in both metrics of vascular reactivity. Among
diabetics, each interquartile range increase in the mean sulfate concentration over the previous 6 days was
associated with a 5.4% decrease in nitroglycerin-mediated reactivity (95% CI: -10.5 to -0.1) and flow-
mediated reactivity (-10.7% [95% CI: -17.3 to -3.5]). Also among diabetics, each interquartile range
increase in the mean PM2 5 concentration over the previous 6 days was associated with a significant 7.6%
decrease in nitroglycerin-mediated reactivity (95% CI: -12.8 to -2.1) and a non-significant 7.6% decrease
in flow-mediated reactivity (95% CI: -14.9 to 0.4). Each interquartile range increase in the mean BC
concentration over the previous 6 days was associated with a 12.6% decrease in flow mediated reactivity
(95% CI: -21.7 to -2.4), but not nitroglycerin-mediated reactivity. PNC was associated with
December 2008
6-36
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
non-significant decreases in both. Effect estimates were larger for type II diabetics than type 1 diabetics
(O'Neill et al, 2005a).
Dales et al. (2007) conducted a panel study of 39 healthy volunteers who sat at 1 of 2 bus stops in
Ottawa, Canada for 2 h (Dales et al., 2007). Flow-mediated vasodilation of the brachial artery was
measured immediately after the bus stop exposure, but not before. The mean PM25 concentrations,
measured at the 2 bus stops, were 40 and 10 |_ig/nr\ Using mixed effects models with random slopes,
adjusting for location, time of the day, and temperature, Dales et al. (2007) examined the association
between flow-mediated vasodilation and 2-h mean PM2 5, PMi, N02, and traffic density at the bus stop
(vehicles/h). The authors report that each 30 (ig/m3 increase in 2-h mean PM2 5 concentration was
associated with a significant 0.48% reduction in flow-mediated dilation (FMD) (p = 0.05). This
represented a 5% relative change in the maximum ability to dilate (Dales et al., 2007).
This same research group conducted a panel study of 25 type I or II diabetic subjects living in
Windsor, Ontario (aged 18-65) (Liu et al., 2007b). For each subject, personal PMi0 concentrations were
measured for 24 h before measurements of BAD, FMD, and other biomarkers. The mean 24-h mean PMi0
concentration, measured with personal monitors, was 25.5 (ig/m3. Each 10 (ig/m3 increase in 24-h mean
PM10 concentration was associated with a 0.20% increase in end-diastolic FMD (95% CI: 0.04 to 0.36)
and a 0.38% increase in end-systolic FMD (95% CI: 0.03 - 0.73), but decreases in end-diastolic basal
diameter (-2.52 (im [95% CI: -8.93 to 3.89]) and end-systolic basal diameter (-9.02 (im [95% CI: -16.04
to -2.00]) (Liu et al., 2007b).
Rundell et al. (2007) examined the change in FMD associated with high and low PM,
(0.02-1.0 (.im) pollution in a panel of 16 young intercollegiate athletes (mean age = 20.5 ± 2.4 years), who
were non-smokers, non-asthmatics, and free of cardiovascular disease (Rundell, 2007). Each subject had
FMD of the brachial artery measured 10-20 mins before and 20-30 mins after each of two 30 min exercise
tests (85-90% of maximal HR). The exercise tests were done outside either on an inner campus location
free of automobile and truck traffic (low PMi; mean = 5309 ± 1942 particles/cm3) or on a soccer field
adjacent to a major highway (high PMi; mean = 143,501 ± 58,565 particles/cm3). The order of the
exercise test locations was chosen randomly. Using paired t-tests for analysis, they reported FMD was
impaired after high PMi exposure (pre-exercise: 6.8% ± 3.58%; post-exercise: 0.30% ± 2.74%;
p = 0.0001 for the change) but not low PMi exposure (pre-exercise: 6.6% ± 4.04%; Post-exercise:
4.89% ± 4.42%; p-value for the change not given). Further, they found basal brachial artery
vasoconstriction (4%; pre-exercise BAD: 4.66 ± 0.61 mm; post-exercise BAD: 4.47 ± 0.63 mm;
p = 0.0002 for the change) after the 'high PMi' exposure, but not the 'low PMi' exposure (-0.3%
pre-exercise BAD: 4.66 ± 0.63 mm; post-exercise BAD: 4.68 ± 0.61 mm; p-value for the change not
given) (Rundell, 2007).
December 2008
6-37
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Each study demonstrated an acute association between measures of vascular function and ambient
PM concentrations (Dales et al., 2007; Liu et al., 2007b; O'Neill et al., 2005a; Rundell, 2007). Two
studied a panel of diabetics (Liu et al., 2007b; O'Neill et al., 2005a), and two a panel of young healthy
subjects (Dales et al., 2007; Rundell, 2007). Only one study investigated multiple lags (lag days 0 to 6)
and reported the strongest association with the 6 day mean PM concentration (O'Neill et al., 2005a). In
other studies, responses were observed in as short as 30 mins after the exposure (Rundell, 2007). The
Rundell et al. (2007) findings are consistent with other studies showing an adverse response to ambient
particulate pollution emitted from vehicular traffic (2007a; Adar et al., 2007b; Riediker et al., 2004a;
Riediker et al., 2004b).
6.2.4.2. Human Clinical Studies
Some evidence of a PM-induced increase in brachial artery vasoconstriction is presented in the
2004 PM AQCD. Brook et al. (2002) exposed 24 healthy adults to PM25 CAPs (150 (.ig/nr1) along with
120 ppm 03 for a period of 2 hours. A significant decrease in BAD was observed immediately following
exposure compared with filtered air control. No significant changes were observed in either
endothelial-dependent or endothelial-independent vasomotor function, as determined by FMD and
nitroglycerin-mediated dilatation, respectively. However, the lack of any FMD in these subjects suggests
that there may have been a technical problem with the measurement of FMD in this study. A subsequent
analysis of the CAPs constituents revealed a significant negative association between the post-exposure
change in BAD and both the organic and EC concentrations of CAPs (Urch et al., 2004). However, the
observed vasomotor effects cannot conclusively be attributed to PM2 5, as subjects were exposed
concurrently to PM25 and 03.
In a more recent randomized cross-over controlled human exposure study, Mills et al. (2005)
exposed 30 healthy men (20-38 years old) to both diluted DE (300 |ig/nr') and filtered air control for 1-h
with intermittent exercise. Half of the subjects underwent vascular assessments at 6 to 8 hours following
exposure to diesel or filtered air, while in the other 15 subjects, vascular assessments were performed at 2
to 4 hours post-exposure. DE attenuated forearm blood flow increase induced by bradykinin,
acetylcholine, and sodium nitroprusside infusion measured 2 and 6 hours after exposure. The authors
postulated that the effect of diesel on vasomotor function may be the result of reduced NO bioavailability
in the vasculature stemming from oxidative stress induced by the nanoparticulate fraction of DE. A
diesel-induced decrease in the release of t-PA was also observed at 6 hour post-exposure, which may
provide additional mechanistic evidence supporting the observed association between air pollution and
MI.
December 2008
6-38
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
To further investigate the effects of DE on vasomotor function, Mills et al. (2007) exposed 20 men
(average age 60 years) with previous MI on two separate occasions to dilute DE (300 |ig/m3: mean
particle size 54 nm) or filtered air for 1 h with intermittent exercise. Contrary to previous findings in
younger, healthy adults (Mills et al., 2005), DE was found not to affect vasomotor function in peripheral
resistance vessels at 6 h post-exposure as measured by endothelium-dependent (acetylcholine) and
endothelium-independent (sodium nitroprusside) vasodilation (forearm blood flow). However, similarly
to younger adults, bradykinin-induced release of t-PA was observed to decrease 6 hours following
exposure to DE in older adults with stable coronary artery disease. In a subsequent randomized crossover
study, Mills et al. (2008) evaluated the effect of fine and ultrafine CAPs on vasomotor function in a group
of 12 males with stable coronary heart disease (mean age = 59 years), as well as in 12 healthy males
(mean age = 54 years). Relative to filtered air exposure, exposure to PM (average
concentration =190 |ig/m3) did not significantly affect vascular function in either group. The authors
attributed the lack of response in endothelial function to the composition of the CAPs used in the study,
which were low in combustion-derived particles and consisted largely of sea salt.
Peretz et al. (2008b) exposed both healthy adults (n = 10) and adults with metabolic syndrome
(n = 27) for 2 h to filtered air and two concentrations of diluted DE (fine PM concentrations of 100 and
200 |ig/m3). Compared with filtered air, DE at 200 |ig/m3 elicited a statistically significant decrease in
BAD (0.11 mm; 95% confidence interval, 0.02-0.18 mm) immediately following exposure. A smaller
diesel-induced decrease in BAD (0.05 mm) was observed following exposure to 100 |ig/nr\ Plasma levels
of endothelin-1 (ET-1) were observed to increase relative to filtered air exposure approximately 1 h after
exposure to 200 |ig/m3 DE (p = 0.01). Exposure to DE was not shown to significantly affect
endothelium-dependent flow-mediated dilatation. The results of this study provide evidence of an acute
endothelial response and arterial vasoconstriction resulting from short-term exposure to DE.
Diesel-induced changes in vasoconstriction and ET-1 release were more pronounced in the healthy
subjects than in the subjects with metabolic syndrome. The authors attributed this to small number of
normal individuals, and also postulated that subjects with metabolic syndrome may have stiffer vessels
that are not as responsive to vasoconstrictor stimuli.
Tornqvist et al. (2007) evaluated diesel-induced changes in vascular function 24 h following a 1-h
exposure with intermittent exercise in a group of 15 healthy male subjects (18-38 years old). Subjects
were exposed to DE at a particle concentration of 300 |ig/m3. as well as filtered air in a randomized
cross-over study design. Compared with filtered air, exposure to DE significantly reduced
endothelium-dependent (acetylcholine) vasodilation (p = 0.01) at 24-h post exposure. Bradykinin-induced
vasodilation was marginally attenuated by DE (p = 0.08), while no effects of diesel on
endothelium-independent vasodilation (sodium nitroprusside) were observed. Although the release of t-PA
December 2008
6-39
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
was not affected by DE 24 h following exposure, the authors suggest that the persistent association
between diesel exposure and vasomotor function observed in this study provides supporting mechanistic
evidence of the observation of an increase in cardiovascular events occurring 24 h after a peak in PM
concentration.
In the previously described studies by Mills et al. (2005; 2007), Peretz et al. (2008b) and Tornqvist
et al. (2007), subjects were exposed to DE, which, in addition to PM, includes DE gases such as nitrogen
oxides and carbon monoxide. Therefore, it is possible that the observed effects may be due in part to
exposure to non-particle components of DE. However, Brauner et al. (2008) found that using high
efficiency particle air (HEPA) filters in the homes of 21 healthy older couples (60-75 years old)
significantly improved microvascular function (assessed by digital peripheral artery tone after arm
ischemia) at the end of a 48-h period when compared against a 48 hour period in the same group without
filtration. In this study, the HEPA filter reduced average PM2.5 concentration from 12.6 to 4.7 |ig/m3: no
difference in N02 concentration was observed between the two exposure periods. The effect of particulate
exposure on microvascular function demonstrated in this study can most likely be attributed to changes in
autonomic function as no changes in inflammatory mediators, oxidative stress parameters, or coagulation
markers were observed.
A recent study evaluated the effect of EC UFP on vascular function in a group of sixteen healthy
subjects (18-40 years old) (Shah et al., 2008). Subjects were exposed via mouthpiece to 50 (.ig/nr1 UFP for
2 h with intermittent exercise. Venous occlusion plethysmography was used to measure reactive
hyperemia of the forearm prior to exposure, immediately following exposure, and 3.5 h, 21 h, and 45 h
post-exposure. Peak forearm blood flow was observed to increase with air at 3.5 hours post-exposure, but
not following exposure to UFP (p = 0.03). Venous nitrate levels were significantly lower at 21 h following
exposure to UFP compared with air (p = 0.038). Based on these findings the authors concluded that UFP
may induce vasomotor dysfunction and reduce NO bioavailability.
Taken together, the two studies by Mills et al. (2005; 2007) along with the studies by Peretz et al.
(2008a) and Tornqvist et al. (2007) suggest that, in healthy subjects, DE exposure inhibits
endothelium-dependent and endothelium-independent vasodilation acutely (within 2-6 h), and that the
suppression of endothelium-dependent vasodilation may remain up to 24 h following exposure. In
patients with coronary artery disease, vasodilator function does not appear to be affected 6-8 h following
exposure; however, vascular assessments were not performed at earlier time points. In addition, the use of
medications in these patients may have blunted the response to PM. The findings of Shah et al. (2008)
suggest that UFP carbon core may be sufficient to produce small changes in systemic vascular function,
but the mechanisms remain obscure. The authors demonstrated a decrease in nitrate levels following
December 2008
6-40
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
exposure to EC UFP; however, venous nitrite level, which more closely reflects NO production, was
unchanged.
6.2.4.3. Toxicological Studies
Vascular dysfunction is a function of altered production of vasoconstrictors and vasodilators. In the
2004 PM AQCD, studies examining ET as an activator of vasoconstriction were limited to those
conducted by Bouthiller et al. (1998) and Vincent et al. (2001), in which increased plasma ET levels were
observed in rats exposed to high concentrations (40 or 5 mg/m3) of resuspended Ottawa (EHC-93) or
diesel PM, respectively. The authors postulated that PM altered vasoconstriction via elevated ET. No
studies were cited in the 2004 PM AQCD that looked at direct measures of vasoreactivity.
As this area is newly emerging, numerous studies are included below that utilize intratracheal
instillation and/or high exposure levels; the studies that expose vessels directly to particles ex vivo are
included in the annex only, as their relevance is questionable. There is clearly a need for more
toxicological research examining the relationship between vascular measurements and PM exposures
using ambient particles at lower concentrations. Furthermore, no new studies have advanced the
knowledge in regards to ET as a biomarker of PM-induced vasoconstriction since the last PM review.
Nurkiewicz et al. (2004; 2006) have shown a relationship between the impairment of
endothelium-dependent dilation in the systemic microvasculature following intratracheal instilled PM.
Using a model of instilled residual oil fly ash (ROFA; partially soluble) or Ti02 (insoluble) particles (0.1
or 0.25 mg/rat), the authors studied the systemic microvascular effects (right spinotrapezius muscle) in
healthy Sprague Dawley rats 24 h after exposure. The authors found comparable dose-dependent
impairment of calcium ionophore NO dependent-induced dilation in the arteriole, regardless of the
particle type (Nurkiewicz et al., 2004). NO-independent arteriolar dilation (measured by inhibiting local
NO synthesis with NG-monomethyl-L-arginine) was also impaired by ROFA, but a NO donor (sodium
nitroprusside) did not affect dilation when directly applied to the exterior arteriolar wall; this indicates
that arteriolar smooth muscle responsiveness was unaltered with ROFA exposure. Arteriole adrenergic
sensitivity to phenylephrine was not affected by 0.25 mg ROFA, indicating that contractile activity was
unchanged (Nurkiewicz et al., 2006).
Identification of increased adherence and "rolling" of leukocytes in the systemic vasculature was
further characterized in a follow-up intratracheal instillation study, indicative of an activated endothelium,
which was induced similarly for both ROFA and Ti02 (Nurkiewicz et al., 2006). Vascular deposition of
myleoperoxidase (MPO) was observed in the spinotrapezius muscle and the authors suggested that the
adherent leukocytes may have deposited the MPO to be taken up by endothelial cells (Nurkiewicz et al.,
2006). However, this is in contrast to another study (Cozzi et al., 2006) that did not find changes in blood
December 2008
6-41
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
neutrophil MPO release in mice exposed to ultrafine PM via intratracheal instillation (100 |ig: Chapel
Hill, NC; ICR mice; assessed 24-h post-exposure), although this finding may be a reflection of differing
analysis approaches. Lastly, increased oxidative stress in the arteriolar wall was increased following 0.25
mg ROFA, as measured by the tetranitroblue tetrazolium reduction method. Ti02 and ROFA induced
varying degrees of pulmonary inflammation in these animals, but elicited very similar vascular effects,
indicating that the vascular responses may be due to PM presence in the lung rather than its intrinsic
pulmonary toxicity. Furthermore, the microvasculature responses do not appear dependent upon the
soluble PM components.
A subsequent study by Nurkiewicz et al. (2008) compared the arteriole dilation responses with
inhalation exposure to fine or ultrafine Ti02 (1 |im and 21 nm, respectively; mean mass concentration
range 3-16 and 1.5-12 mg/m3, respectively) for durations ranging from 4 to 12-h in Sprague Dawley rats.
Similar to what was observed in the previous studies, both size fractions of Ti02 induced impaired
dilation with calcium ionophore infusion in a dose-dependent manner. When ultrafine and fine Ti02 were
compared at similar mass doses, the systemic microvascular dysfunction was greater with the ultrafine
particles. Furthermore, two exposures of differing durations and concentrations that produced equal
calculated pulmonary deposition of ultrafine Ti02 (30 |ig) demonstrated similar dilation responses,
indicating that impairment is dependent upon the time x concentration product.
Tamagawa et al. (2008) reported reduced acetylcholine (ACh)-stimulated relaxation in carotid
arteries from rabbits (New Zealand Whites) exposed to PM (EHC-93) for 1 or 4 weeks (2.6 mg/kg on 1st,
3rd, and 5th days or 2 mg/kg twice weekly, respectively) via intrapharyngeal instillation.
Endothelium-dependent NO-mediated vasorelaxation correlated with increased serum IL-6 levels in the
acute study and during weeks 1 and 2 of the 4-week exposure, which may indicate a role for systemic
inflammation in the response. Maximal SNP-induced dilation was not affected by PM exposure,
indicating that the dilatory response was not acting via endothelium-independent NO-mediated
mechanisms. This finding is consistent with that by Nurkiewicz et al. (2004) and suggests that the
arteriolar smooth muscle is not involved in the PM-impaired dilatation response.
Cozzi et al. (2006) used ICR mice to examine the effects of ultrafine PM exposure on vascular
reactivity following PM exposure and ischemia/reperfusion injury. Mice received 100 |ig of ultrafine PM
collected from Chapel Hill, NC via intratracheal instillation. At 24-h following exposure, mice were given
an experimental ischemia/reperfusion injury by temporarily occluding the left anterior descending artery
for 20 mins. Reperfusion of the left anterior descending artery lasted 2-h and aortic rings were evaluated
for their contractile and dilatory responses. Maximum ACh-induced relaxation was impaired in
PM-exposed vessels, as well as a rightward shift in sensitivity to ACh. There was no difference in
constriction to phenylephrine between aortic rings from control and PM-exposed mice. The reduced
December 2008
6-42
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
ACh-induced relaxation may be important for reperfiision of critical vascular beds following occlusion,
potentially leading to a greater area of infarction (as in this study). A new study in dogs supports the
results observed in the above study and provides evidence of reduced myocardial blood flow following
PM exposure (Bartoli et al., 2008), and is discussed in more detail in Section 6.2.3.3.
Vasoreactivity of aortic rings was measured 4 and 24-h following intratracheal instillation exposure
to 10 mg/kg PM (EHC-93; SH rats), with an increase in ACh-induced vasorelaxation observed with PM
exposure (Bagate et al., 2004a). This endothelium-dependent response was greatest at 4 h and was still
significantly different from the control group at 24 h. Similarly, vasorelaxation induced by SNP
(endothelium-independent) 4-h post-PM exposure was enhanced. Furthermore, the vasorelaxation
response was attenuated after denudation of the aortic rings, suggesting that the effect is
endothelium-dependent. The findings of enhanced dilation with PM exposure contrast with those reported
by Nurkiewicz et al. (2004; 2006), Tamagawa et al. (2008), and Cozzi et al. (2006) and may be
attributable to differences in PM type, animal species, or disease models. The authors attribute their
findings to the SH rat as a well-documented model of sympathetic hyperactivity (increased affinity of
aortic smooth muscle a-adrenergic receptors) that demonstrates upregulation of NO formation and/or
release (Safar et al., 2001). No change in vasoconstriction was observed with PM with either
phenylephrine or potassium chloride.
Consistent with the impaired vasodilatory responses observed in the microvasculature and aortic
rings following PM exposure, Courtois et al. (2008) demonstrated less relaxation to ACh in
intrapulmonary arteries of Wistar rats exposed via intratracheal instillation to a very high dose (5 mg) of
ambient PM (SRM1648). This response was only observed 12-h after PM exposure and not at shorter (6
h) or longer (24 or 72 h) time points. Interestingly, fine Ti02 did not alter ACh-induced relaxation.
Sprague Dawley rats were exposed to PM2 5 CAPs (5 h/day x 3 day; mean mass concentration
range 73.5-733 |ig/m3: mean mass concentration range for 3-day exposure 126.1-481.0 |ig/m3: Boston,
MA; March 1997 to June 1998) then the pulmonary arterial vasculature was evaluated (Batalha et al.,
2002). Some animals were repeatedly exposed to S02 (5 h/day x 5 day/wk x 6 wk) to induce chronic
bronchitis (CB). Morphometric measurements indicated that the pulmonary artery lumen-to-wall (LAV)
ratio was decreased for the both CAPs groups compared to the normal/air group. Furthermore, decreased
LAV ratio in CAPs-exposed animals (regardless of pre-treatment) was significantly associated with
particle mass and composition when the mean concentrations from the second and third exposure days
were used in a univariate linear regression. These results indicate a change in vascular tone following
acute exposure to PM.
The venous circulation plays a prominent role in heart failure exacerbation (Gehlbach and Geppert,
2004). In heart failure, patients are often volume overloaded and are subsequently placed on diuretics to
December 2008
6-43
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
alleviate symptoms of pulmonary congestion and chest pain. Knuckles et al. (2008) hypothesized that if
veins constrict in a manner similar to arteries, then patients with severe heart failure may have temporary
shunting of fluid to the pulmonary circulation, which may elicit signs and symptoms of heart failure.
Using mesenteric vessels from mice (C57BL/6) exposed to whole diesel emissions (350 |ig/m3 x 4 h), the
authors reported a significant enhancement of ET-1 -induced vasoconstriction in veins with a much weaker
responses in arteries. In an ex vivo experiment, venous constriction was blocked by the arginine analog,
L-NAME, which eliminates the feedback NOS activation via endothelial ETB receptors; this is indicative
of impaired or uncoupled eNOS. The authors hypothesized that volatile organic compounds might be
responsible these effects, but no significant effects were observed for acetaldehyde, formaldehyde,
acetone, hexadecane, or pristane.
In addition to studies that look at vascular reactivity, three recent studies have examined plasma ET
levels following exposure to vehicle emissions. Circulating levels of ET-1 (measured 18 h post-exposure)
were elevated in animals exposed to gasoline emissions and filtration of particles did not reduce this
effect (see study details in Section 6.2.2.2.) (Campen et al., 2006). The results of Campen et al. (2006) are
consistent with those observed by Bouthillier et al. (1998) following a very high exposure to EHC-93, but
it is difficult to attribute the effects to PM alone, as Campen et al. (2006) showed that the gaseous
gasoline emissions were required for the ET-1 increase. In contrast, a study of old rats (21 month; Fischer
344) exposed to on-road highway aerosols (number concentration range 0.95-3.13* 105 particles/cm3;
Interstate 90 between Rochester and Buffalo, NY) for 6 h demonstrated decreased plasma ET-2 (18 h
post-exposure) and unchanged levels of ET-1 and ET-3 (Elder et al., 2004a).
One study examined the mRNA expression of ET-1 and the ETA receptor in hearts of Wistar Kyoto
rats exposed to CAPs (1 or 4 days; 4.5 h/day; mean mass concentration range 1000-1900 (.ig/nr3; May
2004, November 2004, September 2005; Yokohoma City, Japan) and reported correlations between
mRNA upregulation and increasing PM cumulative mass collected on chamber filters (Ito et al., 2008).
Furthermore, relative cardiac mRNA expressions of ET-1 and ETA receptor were significantly correlated
with CYP1B1 and HO-1 expression, indicating a possible relationship between ET-1 metabolism or
oxidative stress.
Another plasma indicator of vasomotor tone is asymmetric dimethylarginine (ADMA), which is an
endogenous inhibitor of NOS. Dvonch et al. (2004) assessed levels of ADMA in Brown Norway rats 24-h
following a 3-day PM2 5 CAPs exposure (8 h/day; southwest Detroit, MI; July 2002). CAPs (mean mass
concentration 354 |ig/m3) resulted in increased plasma ADMA compared to air controls, although the
levels reported were well below the 2 (J.M/L range associated with increased CVD risk in humans in
chronic studies. Therefore, the preliminary results identified a new potential biomarker of vascular tone
that had not previously been used in air pollution toxicological studies.
December 2008
6-44
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
The toxicological findings with respect to vascular reactivity are generally in agreement and
demonstrate impaired dilation following PM exposure that is likely endothelium dependent. These effects
have been demonstrated in varying vessels (right spinotrapezius muscle, carotid arteries, and aortic rings)
and in response to different PM types (ROFA, Ti02, EHC-93, ultrafine ambient PM). Only one study
showed enhanced dilation with PM exposure, but the authors attributed the conflicting results to the SH
rat. No constriction changes in response to phenylephrine were observed following PM exposure. The
responses observed in the pulmonary circulation after PM exposure include pulmonary vasoconstriction,
decreased LAV ratio, and impaired vasodilation in intrapulmonary arteries. These results are consistent
and indicate altered vascular tone. Enhancement of vasoconstriction in mesenteric veins following DE is
the first study of its kind to report on venous circulatory effects.
The ET responses were mixed, with one study demonstrating ET-1 increases after exposure to
gasoline emissions that were particle independent and another reported decreased ET-2, but no change in
ET-1 or ET-3 with on-road highway exposure. Elevated levels of ET-1 and ETA receptor mRNA were
noted in hearts of rats exposed to CAPs. A relatively novel marker, ADMA, was used to evaluate
vasomotor tone in rats and was found to be elevated following exposure to CAPs, although the results are
preliminary and have not been confirmed.
PM Components
In the Batalha et al. (2002) study described above, univariate analyses were conducted that
regressed log LAV on differential exposure concentrations of tracer elements determined using principal
components analysis with verimax rotation. For CAPs exposure (regardless of pre-treatment), CAPs
mass, Si, Pb, sulfate, EC, and OC were all negatively correlated with LAV ratio. Si and sulfate were
negatively correlated with LAV ratio in normal rats and Si and OC were negatively correlated with LAV
ratio in CB rats. When a multivariate analysis was conducted using normal and CB animals, only the
association with Si remained significant. V was not associated with LAV ratio in any analysis.
6.2.5. Blood Pressure
One of the potential outcomes of air pollution-mediated alterations in vascular tone is its impact on
variable BP or hypertension. BP is tightly regulated by autonomic (central and local), cardiac, renal, and
regional vascular homeostatic mechanisms with changes in arterial tone being countered by changes in
cardiac contractility, HR, or fluid volume. The evidence of PM-induced changes in BP presented in the
2004 PM AQCD is limited and inconsistent. Recent epidemiologic, human clinical, and toxicological
studies have similarly reported conflicting results regarding the effect of PM on BP. However, the
December 2008 6-45 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
majority of these studies have evaluated changes in BP at some point following exposure to PM. A
significant increase in diastolic BP was observed in the only human clinical study that evaluated BP
during exposure (concomitant exposure to CAPs and ozone). In addition, evidence from toxicological
studies suggests that the effect of PM on BP may be modified by health status, as PM-induced increases
in BP have been more consistently observed in SH rats.
6.2.5.1. Epidemiologic Studies
Increased BP was associated with PM concentration in two of three studies reviewed in the 2004
PM AQCD. Increases in left ventricular BP (systolic and diastolic) are well established risk factors for
cardiovascular mortality/morbidity (Welin et al., 1993). Changes in HR and BP both reflect changes in
autonomic tone, and have been examined following short-term increases in PM pollution in several recent
studies.
Ibald-Mulli et al. (2004) examined associations between BP (systolic [SBP], diastolic [DBP]) and
ambient PM2 5 concentrations, UFP counts, and ACP counts in a multicity panel study (Amsterdam,
Netherlands; Helsinki, Finland; Erfurt, Germany) of 131 adults with coronary heart disease (Ibald-Mulli
et al., 2004). Although based on the same ULTRA Study (Timonen et al., 2006) with study methods as
described previously in Section 6.2.1.1., the study period was different. They investigated changes in BP
(SBP and DBP) associated with mean PM2 5, UFP, and ACP concentration/counts (lag days 0, 1, and 2, as
well as the 5 day mean) in each city and then generated a pooled estimate across the cities. The median
PM2 5 concentration, median UFP count, and median ACP count for each city are given below in Table 6-
3. Pooled analyses across all 3 cities showed small, but statistically significant decreases in SBP and DBP
associated with different lagged concentrations/counts of each particulate pollutant.
Table 6-3.
Median particulate concentration.
City
Median daily PM2.5
concentration (|jg/m3)
Median daily UFP concentration
(0.01-0.1 |jm; particles/cm3)
Median daily ACP concentration
(0.1-1.0 |jm; particles/cm3)
Amsterdam,
Netherlands
16.9
17,147
1,874
Erfurt, Germany
16.3
19,198
1,492
Helsinki, Finland
10.6
14,886
1,200
Each 10 (ig/m3 increase in the mean PM2 5 concentration over the previous 5 days was associated
with a 0.36 mmHg decrease in SBP (95% CI: -0.99, 0.27) and a 0.39 mmHg decrease in DBP (95%
CI: -0.75 to -0.03). Each 10,000 particles/cm3 increase in UFP was associated with a 0.72 mmHg decrease
December 2008
6-46
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
in SBP (95% CI: -1.92, 0.49), and a 0.70 mmHg decrease in DBP (95% CI: -1.38 to -0.02). Each 1000
particles/cm3 increase in 5 day average ACP was associated with a 1.11 mmHg decrease in SBP (95%
CI: -2.12 to -0.09) and a 0.95 mmHg decrease in DBP (95% CI: -1.53 to -0.37). The authors concluded
that these findings do not support previous findings of an increase in BP associated with increases in
particulate pollutant concentrations (Ibald-Mulli et al., 2004).
Single city studies examining the association between BP and particulate air pollution have been
done in several U.S. and Canadian cities. Dales et al. (2007) conducted a panel study of 39 healthy
volunteers who sat outside at two different bus stops for 2-h in Ottawa, Canada (Dales et al., 2007). The
median PM2.5 concentrations measured at the bus stops during each 2-h exposure session were 40 and
10 (ig/m3. Post exposure SBP and DBP were not associated with the mean PM2 5 concentration measured
at the bus stops during the 2 hour exposure session. The change in BP (post-exposure - pre-exposure)
could not be evaluated, as health measurements were only made after the 2-h exposure session (Dales et
al., 2007).
Jansen et al. (2005) studied changes in BP among 16 older subjects (aged 60-86 years) with asthma
or COPD in Seattle Washington, associated with indoor, outdoor, and personal PMi0, PM2.5, and BC
measurements (levels within the health measurement session) on 12 consecutive days. The mean daily
outdoor PM10 and PM25 concentrations were 13.47 and 10.47 (ig/m3, respectively. The mean daily
outdoor BC concentration was 2.01 (ig/m3. The study authors reported that no associations were observed
between BP and daily mean PM10, PM2 5, or BC concentrations, but did not present any of these results in
the paper (Jansen et al., 2005).
Zanobetti et al. (2004) examined the association between BP (SBP, DBP, and mean arterial BP) and
mean PM2 5 concentrations in the previous 24, 48, 72, 96, and 120 h in 62 elderly, cardiac rehabilitation
patients in Boston, MA (Zanobetti et al., 2004). The median PM2 5 concentration during the study was
8.8 (ig/m3. Each 10.4 (ig/m3 increase in mean PM2 5 concentration in the previous 120 hours was
associated with significant increases in resting DBP (2.82 mmHg, 95% CI: 1.26, 4.41), SBP (2.68 mmHg,
95% CI: 0.04, 5.38), and mean arterial BP (2.76 mmHg, 95% CI: 1.07, 4.48) (Zanobetti et al., 2004).
Mar et al. (2005b) studied this same PM2 5-BP association in 88 subjects aged >57 years in Seattle,
WA. Among healthy subjects taking medications (bronchodilators, inhale corticosteroids,
anti-hypertensives, beta-blockers, calcium channel blockers, and/or cardiac glycosides), each 10 (ig/m3
increase in mean outdoor PM2 5 concentration on the same day as the BP measurement was made was
associated with small increases in SBP and DBP (the results for these analyses were presented in figures
only). However, among all subjects, each 10 (ig/m3 increase in same day mean PM2 5 concentration was
associated with non-significant decreases in SBP (-0.81 mmHg, 95% CI: -2.34, 0.73) and DBP (-0.46
mmHg, 95% CI: -1.49, 0.57) (Mar et al., 2005b).
December 2008
6-47
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
As described earlier, Ebelt et al. (2005) conducted a repeated measures panel study of 16 patients
with COPD in the summer of 1998 in Vancouver, British Columbia to evaluate the relative impact of
ambient and non-ambient exposures to PM2 5, PM10, and PM10.2.5 on multiple health outcomes including
ectopy and BP. Using the same analytic methods, pollutant concentrations, and lags, they reported
decreased systolic BP associated with same day ambient exposures to each PM size fraction (results were
presented in figures only) (Ebelt et al., 2005).
Two similar studies were done in Incheon, South Korea (Choi et al., 2007) and Taipei, Taiwan
(Chuang et al., 2005b). Both reported significant increases in BP associated with acute increases in
ambient PM. Choi et al. (2007) reported significantly increased SBP and DBP associated with the mean
PM10 concentration over the same and previous 2 days in the warm season only (July to September).
Chuang et al. (2005a) reported significant increases in SBP and DBP associated with the mean UFP count
(0.01 to 0.1 |_im particles) 1 to 3 hours before the BP measurement (Chuang et al., 2005b).
These studies (Choi et al., 2007; Chuang et al., 2005a; Dales et al., 2007; Ibald-Mulli et al., 2004;
Mar et al., 2005b; Zanobetti et al., 2004) are not entirely consistent with regard to their BP-PM
associations. Most have reported increases in SBP and DBP associated with increases in either PM2 5,
PM10, or UFP (Choi et al., 2007; Chuang et al., 2005a; Mar et al., 2005b; Zanobetti et al., 2004).
However, two studies reported small decreases in BP associated with multiple particulate pollutants
(Ibald-Mulli et al., 2004); Mar et al., 2005), Dale et al. (2007) reported no change in BP associated with a
2 hour exposure to bus stop PM2 5 and Jansen at al. (2005) reported null findings among older adults in
Seattle, WA. Exposure lags ranging from 1-3 hours (Chuang et al., 2005a), to the same day (Ebelt et al.,
2005; Mar et al., 2005b; Rich et al., 2008), to the mean across the previous 5 days (Zanobetti et al., 2004)
were reported as having the strongest associations with BP.
Right Ventricular Pressure
Several recent studies, summarized in the section on hospital admissions and ED visits for CVD
causes, have reported increased risk of hospital admissions for congestive heart failure associated with
increased PM concentration on the same day (Wellenius et al., 2005b; 2006b). As a possible mechanism
for these reported associations, Rich et al. (2008) hypothesized that these hospital admissions for
decompensation of heart failure would be preceded by more subtle increases in pulmonary arterial (PA)
and right ventricular (RV) diastolic pressures. They used passively monitored PA and RV pressures on
5807 person-days, among 11 subjects implanted with the Chronicle Implantable Hemodynamic Monitor
[Medtronic, Inc. Medtronic, MN]). Using a two-stage modeling process (generalized additive model and
mixed effects model adjusted for time trend, weekday, calendar month, apparent temperature, and
barometric pressure), they examined the change in daily mean right heart pressures associated with mean
December 2008
6-48
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
PM2 5 concentration on the same and previous 6 days. Each 11.62 (ig/m3 increase in same day mean PM25
concentration was associated with small but statistically significant increases in estimated PA diastolic
pressure (0.19 mmHg; 95% CI: 0.05, 0.33) and RV diastolic pressure (0.23 mmHg; 95% CI: 0.11, 0.34).
These effects were not attenuated when controlling for all lags simultaneously. Thus, PM induced right
heart pressure increases may mark another potential pathway between PM exposure and incidence of
cardiovascular events, but further studies on this same hypothesis are needed for confirmation.
Wellenius et al. (2007) conducted a panel study of 28 subjects living in the greater Boston
metropolitan area, each with chronic stable heart failure and impaired systolic function. They
hypothesized that circulating levels of B-type natriuretic peptide (BNP), measured in whole blood at 0, 6,
and 12 weeks, were associated with acute changes in ambient air pollution, as a possible mechanistic
explanation for the observed association between hospital admissions for heart failure and ambient PM
concentration (2005b; Wellenius et al., 2006b). During the study, the mean PM2 5 concentration was
10.9 (ig/m3, while the mean BC concentration was 0.73 (ig/m3. Using linear mixed models, they reported
no association between any pollutant (PM2 5, CO, S02, N02, 03, and BC) and BNP at any lag (e.g., each
10 (ig/m3 increase in mean daily PM2 5 concentration [0.8% increase in BNP; 95% CI: -16.4, 21.5])
(Wellenius et al., 2007).
6.2.5.2. Human Clinical Studies
Only one controlled human exposure study cited in the 2004 PM AQCD reported any PM-induced
changes in BP. Gong et al. (2003a) found that exposure to PM2 5 (174 (.ig/ni3) decreased systolic BP in
asthmatics, but increased systolic BP in healthy subjects. Among healthy adults, BP was not affected
following 2-h exposures to 200 (.ig/ni3 diesel PM (Nightingale et al., 2000), 150 |ig/nr' PM2 5 CAPs with
120 ppb 03 (Brook et al., 2002), or 10 (.ig/m3 ultrafine carbon particles (Frampton, 2001). One recent
study demonstrated a significant increase (9.3%) in diastolic BP among healthy adults immediately prior
to the end of a 2-h exposure to 150 (.ig/m3 PM2 5 CAPs in combination with 120 ppb 03 (Urch et al.,
2005). The authors also found that the magnitude of change in BP was significantly associated with PM2 5
carbon content, but not total PM2 5 mass. It was postulated that the disparity between these finding and
those of a similar study by the same group (Brook et al., 2002) could be due to differences in
experimental methods. The Brook et al. (2002) study measured post-exposure BP approximately 10 mins
following exposure, while the study by Urch et al. (2005) measured BP during exposure.
The effect of PM on BP has been further investigated in several recent controlled human exposure
studies. Routledge et al. (2006) did not observe any changes in BP among healthy older adults and older
adults with stable angina following a 1-h exposure to UF EC (50 (.ig/ni3). with or without co-exposure to
200 ppb S02. Similarly, Shah et al. (2008) reported no changes in BP among younger healthy adults
December 2008
6-49
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
following exposure to UF EC. Beckett et al. (2005) found no effect of either fine or ultrafine zinc oxide
(500 |ig/m3) on BP in a group of healthy adults. Two new studies have assessed BP changes following a
1-h exposure to DE with a particle concentration of 300 (ig/m3. Tornqvist et al. (2007) reported no
changes in BP following exposure to DE compared with filtered air control; however, post-exposure BP
was only measured 24-h following exposure. In a similar study, Mills et al. (2005) evaluated changes in
BP 2 h following exposure to DE and found a 6 mmHg increase in diastolic BP of marginal statistical
significance (p = 0.08) compared to filtered air control. At lower particle concentrations in dilute DE
(100-200 |ig/m3 fine PM), Peretz et al. (2008a) did not observe any changes in systolic or diastolic BP in
either healthy adults or adults with metabolic syndrome following a 2-hour exposure. The findings of
these new studies do not provide conclusive evidence of an association between PM exposure and an
increase in BP; however, they do suggest that there is a need for additional investigations of PM-induced
changes in BP at various time points following exposure.
6.2.5.3. Toxicological Studies
In healthy animal models, little evidence exists for significant BP changes following inhalation
exposure to environmentally-relevant concentrations of PM. Only one animal toxicological study is
mentioned in the 2004 PM AQCD that examined BP with PM exposure and no effect was observed
(Vincent et al., 2001).
In a recent study of dogs, exposure to PM25 CAPs from Boston (mean concentration 358.1 |ig/m3:
concentration range 94.1-1557 (.ig/nr3) for 5 h resulted in increased systolic BP (2.7 mmHg), diastolic BP
(4.1 mmHg), mean arterial pressure (3.7 mmHg), and lowered pulse pressure (1.7 mmHg) when measured
upstream of the femoral artery (Bartoli et al., 2008). Administration of an a-adrenergic antagonist
(prazosin) prior to CAPs attenuated the BP responses. These findings indicate that CAPs exposure may
have activated a-adrenergic receptors and increased peripheral vascular resistance. Baroreflex sensitivity
was measured immediately before and after exposure during a transient elevation of arterial pressure that
was induced by phenylephrine; increased baroreflex sensitivity was observed in subgroup of dogs
exposed to CAPs, which is consistent with an upregulation of vagal reflexes.
Chang et al. (2004) noted slight increases in SH rat BP (5-10 mmHg) when exposed to ultrafine
CAPs (mean mass concentration 202 (.ig/ni3) during spring months. However, during summer months,
when the CAPs exposure level was less (140 (.ig/ni3), this effect was not observed. It was unclear,
therefore, whether the effects were seasonal or dose-related. In a preliminary study of SH rats exposed to
CAPs during a dust storm event, mean BP was elevated the third and fourth h of a 6-h exposure, although
interpretation of this finding is difficult due to few animals in the exposure group (n = 2; details provided
above) (Chang et al., 2007b). In another study, the increased change in mean BP measured using the tail
December 2008
6-50
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
cuff method following CAPs exposure weakly correlated with PM mass accumulated on chamber filters
over the entire exposure duration (see Section 6.2.4.3. for details) (Ito et al., 2008). Furthermore, ETA
receptor mRNA expression in cardiac tissue was positively correlated with the change in mean BP.
Limited toxicological evidence provides support for elevated BP in dogs or rats with CAPs,
ultrafine CAPs, or CAPs during a dust storm event. However, most of the studies reviewed above were
conducted outside of the U.S.
6.2.6. Cardiac Contractility
The 2004 PM AQCD did not include any toxicological studies that evaluated cardiac contractility
either directly or indirectly following exposure to PM. Two recent animal toxicological studies have
demonstrated reductions in cardiac fractional shortening, diminished ejection shortening, or changes in
the Q A interval following PM exposure. The results of these studies provide some evidence ofPM-
induced changes in cardiac contractility in animal models.
6.2.6.1. Toxicological Studies
The strength of the contracting heart is reflected by its contractility. In heart failure, contractility
wanes significantly and the heart can not compensate during periods of increased physical activity.
Measuring true contractility in a whole animal is difficult, requiring extensive surgical instrumentation
and monitoring. There were no toxicological studies that examined cardiac contractility in the last PM
AQCD.
A recent study using old (18 to 28-mo-old) mice (C57BL/6, C3H/HeJ, and B6C3F1) demonstrated
significant reductions in cardiac fractional shortening (due to increased left ventricular end-diastolic and
end-systolic diameters) following a 4-day (3 h/day) exposure to carbon black (PM2 5 mean concentration
401 |ig/m3: PMio mean concentration 553 (ig/m3) using echocardiography (Tankersley et al., 2008).
Hemodynamic measurements of diminished ejection fraction and maximum change in pressure over time
further supported lowered myocardial contractility. Furthermore, increased right ventricular pressure
associated with elevated right atrial and pulmonary vascular pressures and resistance, was indicative of
pulmonary vasoconstriction in carbon black exposed mice. Heart tissue and isolated cardiomyocytes from
exposed animals demonstrated enhanced ROS that was partially attributable to NOS3-uncoupling and
elevated MMP2 and MMP9 levels, which may implicate myocardial remodeling. The combined results
from this study suggest that cellular mechanisms involving NOS-uncoupled ROS generation likely
mediate PM-induced cardiac effects. Furthermore, mRNA expression for atrial and brain natriuretic
peptides (ANP and BNP, respectively) was increased in hearts from exposed mice compared to control,
December 2008
6-51
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
which is consistent with pulmonary congestion. There were no reported strain-related differences in any
response.
In a less invasive procedure employing radiotelemetry, SH rats were repeatedly and alternately
exposed to ultrafine CAPs in Taiwan on separate days in spring or summer (details provided in Section
6.2.5.3.) (Chang et al., 2004). The QA interval was used as an indirect measure of cardiac contractility
and was calculated as the time duration between the Q wave in the ECG and point A (upstroke in aortic
pressure) in the pressure trace. During the spring exposure, QA interval decreased by 1.6 ms as
demonstrated by fixed effects in linear mixed-effects modeling, which indicates an increase in cardiac
contractility. There were no changes in QA interval observed for the summer months, which may be
attributable to lower ultrafine PM concentrations (mean mass concentration 140 (ig/m3) or differing PM
compositions.
All of the studies above provide some evidence that cardiac contractility may be altered
immediately following PM exposure in animal models. Results from the Tanksersley (2008) study
provide the strongest support for PM-induced contractility changes, as echocardiography and
hemodynamic measurements are well-established for examining cardiac function.
6.2.7. Systemic Inflammation
The evidence presented in the 2004 PM AQCD of increases in markers of systemic inflammation
associated with PM was limited and not sufficient to formulate a definitive conclusion. Recent human
clinical and toxicological studies continue to provide mixed results for an effect of PM on markers of
systemic inflammation including cytokine levels, C-reactive protein, and white blood cell count. While
results from recent epidemiologic studies have also been inconsistent across studies, there is some
evidence to suggest that PM levels may have a greater effect on inflammatory markers among populations
with preexisting diseases.
6.2.7.1. Epidemiologic Studies
Several studies reviewed in the 2004 PM AQCD investigated the association of short-term
fluctuations in PM concentration with markers of inflammation (e.g. oxygen saturation, CRP and white
blood cells). These preliminary studies were found to offer limited support for mechanistic explanations
of the associations between PM concentration and heart disease outcomes. Recent studies, published since
2002, are reviewed below.
Diez-Roux et al. (2006) used the Multi-Ethnic Study of Atherosclerosis (MESA) cohort to examine
the whether C-reactive protein (CRP) increased in response to changes in the mean ambient PM2 5
December 2008
6-52
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
concentrations in the prior day, prior 2 days, prior week, prior 30 days, and prior 60 days (Diez Roux et
al., 2006). Subjects (n = 5,634) lived in either Baltimore City or County, MD, Chicago, IL, Forsyth
County, NC, Los Angeles County, CA, Northern Manhattan and the Bronx, NY, or St. Paul, MN. Each
10 |_ig/m3 increase in the mean PM25 concentrations over the previous 30 days was weakly associated
with a 3% increase in CRP (95% CI: -2% to 10%). Similarly, each 10 (ig/m3 increase in the mean PM2.5
concentrations over the previous 60 days was weakly associated with a 4% increase in CRP (95%
CI: -3%, 11%). However, there was no association between CRP level and the mean PM2.5 concentrations
on the prior day, prior 2 days, or prior week. The authors state that the PM2 5 averaging process for longer
periods of time (i.e. mean for 30 or 60 days vs. the mean for 1 day) reduces exposure error and hence
increases the ability to detect association with 30 and 60 day means, but not 1 and 2 day means (Diez
Roux et al., 2006).
Ruckerl et al. (2007b) conducted a multi-city longitudinal study to examine whether changes in
markers of inflammation were associated with short-term increases in particulate concentrations (PMi0,
PM2 5, particle number concentration [PNC]) and gaseous pollutant (N02, S02, CO, 03). Study subjects
were MI survivors (n= 1003) living in either Athens, Greece; Augsburg, Germany; Barcelona, Spain;
Helsinki, Finland; Rome, Italy; or Stockholm, Sweden. Repeated measurements of IL-6 and CRP were
made during the study. Fibrinogen was also measured in this study and results are discussed in Section
6.2.8.1. The mean city-specific pollutant concentrations during the study are shown below in Table 6-4.
Table 6-4.
Ambient concentrations in six European cities.
Pollutant
Helsinki
Stockholm
Augsburg
Rome
Barcelona
Athens
PNC (particles/cm3)
8,534
9,748
11,876
34,450
18,133
20,589
PM2.5 (|jg/m3)
8.2
8.8
17.4
24.5
24.2
23.0
PM10 ((jg/rn3)
17.1
17.8
33.1
42.1
40.7
38.5
Source: Ruckerl et al. (2007b)
In pooled analyses, each interquartile range (not provided) increase in PNC in the 12 to 17 hours
before the health measurement was associated with a 2.7% increase in the geometric mean IL-6 (95%
CI: 1.0, 4.6). None of the pollutants, at any lag, were associated with CRP levels in these subjects. There
did not appear to be effect modification of these results by smoking, diabetes, or heart failure (Ruckerl et
al., 2007b). Two smaller studies conducted by the same group of investigators among 57 male patients
with coronary heart disease in Efurt, Germany found associations of UFP, ACP and PMi0 with CRP
(Ruckerl et al., 2006) and UFP and ACP with sCD40L, a marker for platelet activation (Ruckerl et al.,
2007b).
December 2008
6-53
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Single city studies of several markers of inflammation have also been conducted in the U.S. and
Canada. To examine changes in inflammation related to short-term fluctuations in air pollution, Delfino et
al. (2008) measured CRP, IL-6, TNF-a, sP-selectin, sVCAM-1 and sICAM-1 in blood during a period of
12 weeks. Associations of these markers with average PM concentration (quasi-ultrafine [PM0 25],
PM0.25-2.5, PM10-2 5, EC, OC, BC, primary OC, secondary OC, PN) 24-h to 9 days prior to the blood draw
were examined. Subjects included residents of two downtown Los Angeles nursing homes who were 65+
years old with a history of coronary artery disease. Both 24-h average and multiday average
concentrations of quasi-ultrafine, EC, primary OC, BC, PN and gaseous pollutants were associated with
CRP, IL-6 and sP-selectin in this study.
Pope et al. (2004a) conducted a panel study of 88 non-smoking, elderly subjects residing in the Salt
Lake City, Ogden, and Provo metropolitan area of Utah (Pope et al., 2004a). The mean PM2 5
concentration during the study was 18.9 |_ig/nr\ Each 100 (ig/m3 increase in same day mean PM2 5
concentration was associated with a 0.81 mg/dL increase in CRP (95% CI: 0.48, 1.14), but not white
blood cells (WBC). However, when excluding 1 influential subject, each 100 (ig/m3 increase in same day
mean PM2 5 concentration was associated with only a 0.19 mg/dL increase in CRP (95% CI: -0.01, 0.39)
(Pope et al., 2004a). Several markers of coagulation were examined in this study and are discussed in
Section 6.2.8.1.
Zeka et al. (2006b) studied 710 elderly members of the VA Normative Aging Study to examine
changes in systemic markers of inflammation and acute changes in particulate pollutant concentrations in
the previous 48 hours, 1 week, and 2 weeks (Zeka et al., 2006b). The median 48-h PNC was 24,200
particles/cm3, while the median 48-h PM2 5 concentration was 9.39 (ig/m3. The median 48 hour BC
concentration was 0.61 (ig/m3, while the median 48 sulfate concentration was 1.84 (ig/m3. They did not
find consistent or significant associations with any pollutant and CRP. The authors state that the largest
effects were observed for the mean PNC and BC concentration in the previous 4 weeks, but there were no
consistent findings for lagged PM2 5 or sulfates (Zeka et al., 2006b).
O'Neill et al. (2007) conducted a cross-sectional study of 92 Boston residents with type 2 diabetes,
to examine the association between plasma levels of intercellular adhesion molecule (ICAM-1), vascular
adhesion molecule (VCAM-1) and PM concentrations (O'Neill et al., 2007). Results for markers of
coagulation measured in this study are discussed in Section 6.2.8.1. PM2 5, BC, and sulfate concentrations
were measured 0.5 km from the patient exam site. The mean PM2 5 concentration during the study was
11.4 (ig/m3. The mean BC concentration was 1.1 |_ig/m3. and the mean sulfate concentration was
3.0 (ig/m3. For all moving averages examined (1-6 days), increases in mean PM2.5 and BC concentration
were associated with increased ICAM-1 and VCAM-1 concentrations. Each 7.6 |_ig/m3 increase in the
mean PM25 concentration over the previous 6 days was associated with a 11.76 ng/mL increase in
December 2008
6-54
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
VCAM-1 (95% CI: 3.48-20.70), and each 0.6 |_ig/m3 increase in the mean BC concentration over the
previous 6 days was associated with a 27.51 ng/mL increase in VCAM-1 (95% CI: 11.96, 45.21).
However, there were no consistent associations between the mean sulfate concentration at any lag and any
marker (O'Neill et al., 2007).
Sullivan et al. (2007) conducted a panel study of n = 47 subjects (aged >55 years) either with
COPD (n = 23) or without COPD (n = 24) in Seattle, WA (Sullivan et al., 2007). They examined the
association between levels of CRP and mean daily PM2 5 concentration. Results for fibrinogen and
d-Dimer are discussed in Section 6.2.8.1. The median PM25 concentration during the study was
7.7 (ig/m3. They did not find any associations between 24-h mean PM2 5 concentrations and levels of CRP
in individuals with or without COPD. Similarly, in the study by Liu et al. (2007b), conducted in Toronto
Ontario, neither CRP (0.11 (ig/mL; 95% CI: -0.03-0.25) nor TNF-a (0.03 pg/mL; 95% CI: -0.07-0.13)
was associated with 24-h mean PMi0 concentration. However, significant positive associations with
markers of oxidative stress, FMD and BP were found (Liu et al., 2007b) are discussed in Sections 6.2.9.1.,
6.2.4.1., and 6.2.5.1., respectively.
In the St. Louis Bus Study, each 5.4 (ig/m3 increase in the mean PM2 5 concentration over the
previous week was associated with 5.5% increase in WBC (95% CI: 0.10-11) (Dubowsky et al., 2006).
Each 6.1 (ig/m3 increase in the mean PM2 5 concentration over the previous 5 days was associated with a
14% increase in CRP among all subjects (95% CI: -5.4-37), but an 81% increase in CRP (95% CI: 21-
172) among subjects with diabetes, obesity, and/or hypertension. Associations between PM2 5 and IL-6
were only observed among those with diabetes, obesity, and/or or hypertension (Dubowsky et al., 2006).
In another study of in-vehicle PM, each 10 (ig/m3increase in PM2 5 concentration during a work-shift was
associated with decreased lymphocytes, increased mean corpuscular volume, neutrophils, and CRP over
the next 10-14 hours among 9 healthy North Carolina state troopers (Riediker et al., 2004b).
Using another marker of inflammation, DeMeo et al. (2004) estimated the change in oxygen
saturation and mean PM2 5 concentration in the previous 24 h (DeMeo et al., 2004) in a panel of elderly
subjects. They used the same panel of elderly Boston residents (n = 28) and study protocol and analytic
methods (12 weeks of repeated oxygen saturation measurements) as Gold et al. (2005) and Schwartz et al.
(2005b) in studies of ST-segment depression and HRV, respectively. At each clinic visit, subjects had 5
mins each of rest, standing, post-exercise rest, and 20 cycles of paced breathing. The median PM2 5
concentration during the study period was 10.0 (ig/m3 (Schwartz et al., 2005b). DeMeo et al. (2004)
reported that each 13.4 |_ig/m3 increase in the mean PM2 5 concentration in the previous 6 hours was
associated with a 0.2% decrease in oxygen saturation (95% CI: -0.3-0.0) during the baseline rest period.
Each 13.4 (ig/m3increase in mean 6 hour PM2 5 concentration was also associated with a decline in oxygen
saturation during the post-exercise period (-0.2%; 95% CI: -0.3-0.0), and post-exercise paced breathing
December 2008
6-55
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
period (-0.1%; 95% CI: -0.3-0.0), but not during the exercise period. The authors suggest that these
oxygen saturation reductions may result from pulmonary vascular and inflammatory changes (DeMeo et
al., 2004).
International studies of the effect of air pollution on markers of inflammation have been conducted.
In a large cross-sectional study of healthy subjects in Tel Aviv, Steinvil et al. (2008) examined biological
markers of inflammation (CRP and WBC) collected as part of routine health examinations for 3,659
individuals. Associations with air pollutants (including PMi0) measured at local monitoring sites for the
day of the examination and up to 7 days prior were examined. No significant associations were found
between pollutant levels and indications of enhanced inflammation. By contrast, both PMi0 and PM2 5
were associated with increases in hs-CRP in healthy students in Taiwan (Chuang et al., 2007a).
Summary
The most commonly measured marker of inflammation in the studies reviewed was CRP. CRP was
adversely associated with PM in some (Chuang et al., 2008; Delflno et al., 2008; Diez Roux et al., 2006;
Dubowsky et al., 2006; Riediker et al., 2004b) but not all studies (Liu et al., 2007b; Ruckerl et al., 2007a;
Steinvil et al., 2008; Sullivan et al., 2007; Zeka et al., 2006b). A multi-city study of MI survivors in
Europe failed to provide evidence of an effect of PM on CRP (Ruckerl et al., 2007a). However, IL-6 was
associated with PNC 12-17 h prior to the health measurement in this population (Ruckerl et al., 2007a).
Two studies reported an adverse associations with markers of inflammation among diabetics or those with
hypertension (Dubowsky et al., 2006; O'Neill et al., 2007). Several other markers of inflammation have
been examined in relation to PM (e.g. VCAM, ICAM, sP-selectin, oxygen saturation) but the number of
studies examining the same marker is too few to allow results to be compared across studies.
6.2.7.2. Human Clinical Studies
Several human clinical studies were included in the 2004 PM AQCD which evaluated markers of
systemic inflammation following exposure to PM. Salvi et al. (1999) exposed 15 healthy volunteers
(21-28 years old) for 1 h to DE (300 |ig/m3 particle concentration) and observed a significant increase in
neutrophils in peripheral blood 6 hours post-exposure compared with filtered air control. Gong et al.
(2003a) did not observe any effect of PM2 5 CAPs (174 |ig/m3) on serum amyloid A, while Frampton
(2001) reported no change in leukocyte activation following exposure to a low concentration (10 (.ig/nr3)
of UF carbon. In a more recent study, Frampton et al. (2006) evaluated the effect of varying
concentrations (10-50 |ig/m3) of UF carbon on blood leukocyte expression of adhesion molecules in
healthy and asthmatic adults. Healthy subjects (n = 40) were exposed for 2-h to filtered air and UF carbon
under three separate protocols: 10 |ig/m3 at rest (n = 12), 10 and 25 |ig/m3 with intermittent exercise
December 2008
6-56
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
(n = 12), and 50 (.ig/nr3 with intermittent exercise (n = 16). Asthmatics (n = 16) were exposed at a single
concentration (10 |ig/m3) for 2 h with intermittent exercise. Leukocyte expression of surface markers were
quantified using flow cytometry on peripheral venous blood samples collected prior to and immediately
following exposure, as well as at 3.5 and 21 hours post-exposure. Among healthy resting adults, UFP
exposure at a concentration of 10 |ig/m3 had no effect on blood leukocytes. The expression of adhesion
molecules CD54 and CD 18 on monocytes, and CD 18 on PMNs was shown to decrease with UFP
exposure in healthy exercising adults. In exercising asthmatics, expression of CD lib on monocytes and
eosinophils, as well as CD54 on PMNs were reduced following exposure to UFP. In both asthmatics and
healthy adults, a UFP-induced decrease in eosinophils and basophils was observed. Although the clinical
significance of these findings is unclear, the authors concluded that their findings of UFP-induced
changes in leukocyte distribution and expression were consistent with increased retention of leukocytes in
the pulmonary vasculature, which may be due to an increase in pulmonary vasoconstriction.
In an effort to better understand the inflammatory response of exposure to PM, Peretz et al. (2007)
conducted a pilot study in which gene expression in peripheral blood mononuclear cells (PBMCs) of
healthy human volunteers was evaluated following a 2 hour controlled exposure to DE (200 |ig/m3 PM2 5).
Adequate RNA samples for microarray analysis (Affymetrix U133 Plus 2.0) from both pre- and 4 hour
post-exposure to filtered air and DE were available in 4 of the 11 subjects enrolled. The authors found
differential expression of 10 genes involved in the inflammatory response when comparing DE exposure
(8 upregulated, 2 downregulated) and exposure to filtered air. Two participants had available samples
from 20 h post-exposure. At this time point, DE was associated with 4 differentially expressed genes (1
upregulated, 3 downregulated). However, this study is limited by a small sample size with limited
statistical power.
Cytokines
Changes in plasma cytokine levels (e.g., IL-6 and TNF-a) have not been observed in human
clinical studies of exposures to CAPs (Ghio et al., 2003) or zinc oxide (Beckett et al., 2005). Similarly,
Mills et al. (2005) found no effect of DE (300 |ig/m3) on serum IL-6 or TNF-a among healthy adult
volunteers 6 hours after exposure. However, at the same PM concentration of DE, Tornqvist et al. (2007)
observed a significant increase in plasma IL-6 and TNF-a in a group of healthy adults 24 h following
exposure. Although the physiological significance of this finding is unclear, this study does provide
evidence of a mild systemic inflammatory response induced by exposure to DE.
December 2008
6-57
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Ce// cownfs
Two studies have observed increased peripheral basophils in healthy older adults 4 hours following
a 2 hour exposure to PM2 5 CAPs (200 |ig/m3) (Gong et al., 2004b), and increased white blood cell counts
in healthy young adults 24-h after a 2 hour exposure to PM2 5 CAPs (120 (ig/m3) (Ghio et al., 2003).
Frampton et al. (2006) reported decreases in blood monocytes, basophils, and eosinophils 0-21-h
following exposure to ultrafine carbon (10-50 (.ig/nr3) among exercising asthmatics and healthy adults.
However, other recent human clinical studies have found no association between peripheral blood cell
counts and exposure to fine or ultrafine zinc oxide (Beckett et al., 2005), ultrafine carbon (Routledge et
al., 2006), or DE (Mills et al., 2005; Mills et al., 2007; Tornqvist et al., 2007).
C-reactive Protein
Several controlled human exposure studies have measured CRP to evaluate systemic inflammation
following exposure to PM. In these studies, no statistically significant changes in CRP concentrations
have been observed 0-24 h following controlled exposures to ultrafine, fine, or thoracic coarse CAPs
(Ghio et al., 2003; 2007), ultrafine carbon (Routledge et al., 2006), or DE (Carlsten et al., 2007; Mills et
al., 2005; Mills et al., 2007; Tornqvist et al., 2007).
6.2.7.3. Toxicological Studies
There has been limited evidence that hematopoiesis may occur in animals exposed to PM. Two
studies in the 2004 PM AQCD provided support for this effect, with one study measured stimulated
release of PMNs from bone marrow and another examined peripheral blood PMN and blood cell counts;
however, one study did not find associations between CAPs and peripheral blood counts. Thus, it was
concluded that consistent evidence of PM-induced hematopoiesis remained to be demonstrated. However,
in a study of humans exposed to biomass burning during the 1997 Southeast Asian smoke-haze episodes,
PMio demonstrated the best relationship with blood PMN band cell counts expressed as a percentage of
total PMN at lag 0 and 1, indicating a relatively quick response (Tan et al., 2000).
In a study of fine and ultrafine carbon black particles (Wistar rats; 7 h; mean mass concentration
1400 and 1660 |ig/m3 for fine and ultrafine CB, respectively; mean number concentration 3.8/103 and
5.2/104 particles/cm3, respectively), only ultrafine CB induced elevated blood leukocytes at 0 and 48 h
post-exposure compared to the control rats and no effect was observed at 16 h (Gilmour et al., 2004c).
Smith et al. (2006) examined the hematology parameters in Sprague Dawley rats following a 3-day
inhalation exposure (4 h/day) to coal fly ash (mean mass concentration 1400 |ig/m3) and reported
increased blood neutrophils and reduced blood lymphocytes at 36 h but not 18 h post-exposure.
December 2008
6-58
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Consistent with the studies above, other studies that measure complete blood counts at 18-24-h
after PM exposure do not report increases in WBC or blood neutrophils. A 2-day CAPs study employing
SH rats did not report increased WBC 18-20 h post-exposure (see details above) (Kodavanti et al., 2005).
A study utilizing fine and/or ultrafine CAPs demonstrated decreased WBC in SH rats 18 h after a 2-day (6
h/day) nose-only exposure (details provided in Section 6.2.8.3.) (Kooter et al., 2006). The decrease was
largely attributable to lowered neutrophils in the fine CAPs-exposed rats and reduced lymphocytes in the
ultrafine+fine CAPs animals. In another study, blood neutrophils were decreased in SH rats exposed to
ultrafine carbon black for 6 h and no effects were observed in old Fischer 344 rats (details provided
above) (Elder et al., 2004b).
Elevated systemic IL-6 and TNF-a cytokine levels were observed following PMi0 instillation in
mice (details provided in Section 6.2.8.3.) (Mutlu et al., 2007). IL-6 was decreased with PM exposure in
macrophage-depleted mice, indicating that some of the IL-6 release originated in macrophages. For mice
(male C57B1/6J) exposed to PMi0.2.5 derived from coal fly ash via intratracheal instillation (200 jxg),
increased plasma IL-6 levels were only observed in animals that also received 100 |ig of LPS (Finnerty et
al., 2007); this response was not observed with LPS alone, indicating a role for PM10-2.5. In contrast, an
inhalation study of carbon black in rats did not demonstrate any change in plasma IL-6 levels (details
provided in Section 6.2.4.3.) (Elder et al., 2004b).
Overall, these studies provide evidence of time-dependent responses of systemic inflammation
induced by PM exposure. Alterations in WBC have been reported generally as elevations immediately (0
h) or < 36 h post-exposure and no change or reductions are noted from 18-24 h.
6.2.8. Blood Coagulation
The 2004 PM AQCD presented limited and inconsistent evidence from epidemiologic, human
clinical, and toxicological studies of PM-induced changes in blood coagulation markers. The body of
scientific literature investigating hemostatic effects of PM has grown significantly since the publication of
the 2004 PM AQCD, with epidemiologic studies demonstrating consistent increases in von Willebrand
factor (vWf) associated with PM. Recent human clinical and toxicological studies have also observed
changes in blood coagulation markers (e.g., fibrinogen, vWf, factor VII, t-PA) following exposure to PM.
However, the findings of these studies are somewhat inconsistent, which may be due in part to differences
in the post-exposure timing of the assessment.
December 2008
6-59
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
6.2.8.1. Epidemiologic Studies
Several studies investigating the association of short-term fluctuations in PM concentration with
markers of coagulation (e.g. blood viscosity and fibrinogen) were included in the 2004 PM AQCD. These
preliminary studies were found to offer limited support for mechanistic explanations of the associations of
PM concentration with heart disease outcomes. New studies, published since 2002, are reviewed in this
section.
Liao et al. (2005) used a cross-sectional study to examine the association between short-term
increases in air pollutant concentrations (mean PMi0, N02, CO, S02, and 03 over the previous 3 days) and
several plasma hemostatic markers (fibrinogen, factor VIII-C, vWF, albumin). Study subjects were middle
aged participants in the ARIC (Atherosclerosis Risk in Communities) study (n = 10,208), and were
residents of Washington County, MD, Forsyth County, NC, selected suburbs of Minneapolis, MN, or
Jackson, MS. The mean PMi0 concentration during the study was 29.9 (ig/m3. Liao et al. (2005) found
that each 12.8 (ig/m3 increase in the mean PMi0 concentration 1 day before the health measurements were
made was associated with a 3.93% increase in vWF (95% CI: 0.40-7.46) among diabetics, but not among
non-diabetics (-0.54%, 95% CI: -1.68-0.60). Each 12.8 (ig/m3 increase in the mean PMi0 concentration 1
day before the health measurements were made was also associated with a 0.006 g/dL decrease in serum
albumin (95% CI: -0.012 to 0.000) among those with CVD, but not among those without CVD (0.029
g/dL increase, 95% CI: -0.004 to 0.062). The mean CO concentration on the previous day was also
associated with a significant decrease in serum albumin. The authors reported significant curvilinear
associations between PM10 and factor VIII-C, and suggest that this may indicate a threshold effect.
Similar curvilinear associations were observed between 03 with fibrinogen, and vWF, and S02 with factor
VIII-C, WBC, and serum albumin (Liao et al., 2005). Liao et al. (2005) did not observe a significant
association with fibrinogen. However, in the European multicity study described in Section 6.2.7.1.,
Ruckerl et al. (2007b), found that each 13.5 (ig/m3 increase in the mean PMi0 concentration over the
previous 5 days was associated with a 0.6% increase in the arithmetic mean fibrinogen level (95%
CI: 0.1-1.1.
Similar studies have been done in other U.S. and Canadian cities. To examine changes in
inflammation related to short-term fluctuations in air pollution, Delfino et al. (2008) measured fibrinogen
and D-dimer in blood of subjects who resided at two downtown Los Angeles nursing homes. As described
in Section 6.2.7.1., measurements were made over a period of 12 weeks and subjects were 65+ years old
with a history of coronary artery disease. These markers were not associated with the broad array PM
metrics studied (e.g. quasi-ultrafine, PM0.25-2.5, PM10-2.5, EC, OC, primary OC, BC). In the study of 92
Boston residents with type 2 diabetes described previously, O'Neill et al. (2007) found that increases in
mean PM2 5 and BC concentration were associated with vWF concentrations for all moving averages
December 2008
6-60
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
examined (1-6 days). Reidiker et al. (2004b) also reports that in-vehicle PM25 was associated with
decreased vWF over the next 10-14 hours among 9 police troopers. However, in the panel study described
previously in Section 6.3.7.1, Sullivan et al. (2007) did not observe associations with fibrinogen, or
D-dimer in individuals with or without COPD (Sullivan et al., 2007). However, neither RBC, platelets nor
blood viscosity were associated with PM2 5 concentration in a panel study of 88 non-smoking elderly
subjects residing in the Salt Lake City, Ogden and Provo metropolitan area of Utah (Pope et al., 2004a).
Although Zeka et al. (2006b) did not observe an association with CRP in the analysis of the NAS
population in Boston (Section 6.3.7.1), increased fibrinogen level was associated with increases in the the
number of particles/cm3 over the previous 48 hours and 1 week, and an incremental increase in BC
concentration over the previous 4 weeks (Zeka et al., 2006b). There were no consistent findings for
lagged PM2 5 or sulfates (Zeka et al., 2006b).
Three studies of coagulation markers were conducted outside the U.S. and Canada. In a study of
healthy individuals in Taiwan, adverse associations were also observed between between PM2 5, PMi0,
nitrate, and sulfate concentrations, fibrinogen and plasminogen activator fibrinogen inhibitor-1 (PAI-1)
(Chuang et al., 2007a). A large cross-sectional study of healthy subjects in Tel-Aviv, Steinvil et al. (2008)
examined fibrinogen collected as part of routine health examinations for 3,659 individuals. No significant
associations were found between pollutant levels (lagged 1-7 days) and fibrinogen. Finally, Baccarelli and
colleagues reported adverse associations between PM10 and prothrombin time (Baccarelli et al., 2007a).
Summary
The most commonly measured markers of coagulation in the studies reviewed were fibrinogen and
vWF. Most studies reported effects of PM on vWF (Liao et al., 2005; 2007; 2004b). Results for fibrinogen
were less consistent (2008; Liao et al., 2005; Ruckerl et al., 2006; Sullivan et al., 2007; Zeka et al.,
2006b) Positive associations with fibrinogen were reported in older adults residing in Boston (Zeka et al.,
2006b) and in the multicity European study of MI survivors. Several other markers have been examined
(e.g. D-dimer, prothrombin time) but not in adequate numbers of studies to allow comparisons across
studies.
6.2.8.2. Human Clinical Studies
In two separate studies conducted by Ghio and colleagues, controlled exposures (2 hours) to fine
CAPs (Chapel Hill, NC) at concentrations between 15 and 350 (ig/m3 have been shown to increase blood
fibrinogen 18- to 24-h following exposure among healthy adults (Ghio et al., 2000; 2003). Increases in
blood fibrinogen or factor VII would suggest an increase in blood coagulability, which could result in an
increased risk of coronary thrombosis. However, a similar study conducted in Los Angeles observed a
December 2008
6-61
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
PM2 5 CAPs-induced decrease in factor VII blood levels in healthy subjects and found no association
between PM2 5 CAPs and blood fibrinogen among healthy and asthmatic volunteers (Gong et al., 2003a).
Since the publication of the 2004 PM AQCD, several new controlled human exposure studies have
evaluated the effects of PM on blood coagulation markers. Routledge et al. (2006) did not observe any
changes in fibrinogen or D-dimers following a 1-h exposure to ultrafine carbon among a group of resting
healthy older adults and older adults with stable angina. Similarly, Beckett et al. (2005) found no changes
in hemostatic markers (e.g., factor VII, fibrinogen, and von Willebrand factor) following exposure to
ultrafine and fine zinc oxide. Mills and colleagues have recently demonstrated a significant effect of DE
(300 |ig/m3) on fibrinolytic function both in healthy men and in men with coronary heart disease (Mills et
al., 2005; 2007). In both groups of volunteers, bradykinin-induced release of tissue plasminogen activator
(tPA) was observed to decrease 6 hours following exposure to DE compared to filtered air exposure. The
same group did not observe an attenuation of t-PA release 24-h after a 1-h exposure (300 |ig/m3) to DE in
a group of health adults (Tornqvist et al., 2007). Carlsten et al. (2007) conducted a similar study involving
exposure of healthy adults to DE with a particulate concentration of 200 |ig/nr\ Although the authors
observed an increase in D-dimer, von Willebrand factor, and platelet count 6 hours following exposure to
DE, these increases did not reach statistical significance. In a subsequent study with a similar study
design, the same researchers found no effect of a 2-h exposure to DE (100 and 200 |ig/m3 PM2 5) on
prothrombotic markers in a group (n = 16) of adults with metabolic syndrome (Carlsten et al., 2008). The
authors postulated that the lack of significant findings could be due to a relatively small sample size. In
addition, Carlsten et al. (2007; 2008) exposed subjects at rest while Mills et al. (2005) exposed subjects to
a higher concentration (300 |ig/m3) with intermittent exercise.
Barregard et al. (2006) recently evaluated the effect of WS on markers of coagulation,
inflammation, and lipid peroxidation. Subjects (n = 13) were healthy males and females (20-56 years old)
and were exposed for 4 hours to PM25 concentrations of 240-280 |ig/nr\ The authors reported an increase
in serum amyloid A at 0, 3, and 20 h following exposure to WS, as well as an WS-induced increase in the
ratio of factor VHI/von Willebrand factor, which is an indicator of an increased risk of venous
thromboembolism. Samet et al. (2007) reported an association between various coagulation markers and
exposure to ultrafine, fine, and thoracic coarse CAPs among healthy adults (18-40 years old). Results
from exposures to ultrafine and thoracic coarse CAPs are presented in summary form, while the
relationship between fine CAPs and an increase in blood fibrinogen has been described in detail in Ghio
et al. (2000). Exposure to thoracic coarse CAPs did not result in a statistically significant change in
prothrombin, fibrinogen, factor VII, t-PA, D-dimer, or von Willebrand factor. However, the authors
observed a trend in CAPs-induced levels of several coagulation factors which they suggested could be
viewed as "pro-clotting." Exposure to ultrafine CAPs was associated with a significant increase (17%) in
December 2008
6-62
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
the concentration of D-dimer. Whereas many coagulation markers provide evidence of an increased
potential to form clots (e.g., an increase in fibrinogen or a decrease in t-PA), D-dimer is actually a
degradation product of a blood clot that has formed. Therefore, the preliminary finding of an association
between ultrafine CAPs and elevated levels of D-dimer is potentially very important to our understanding
of the relationship between elevated concentrations of PM and cardiovascular morbidity and mortality
observed in epidemiologic studies. Taken together, these new studies have provided additional evidence
that short-term exposure to PM at near ambient levels may have small, yet statistically significant effects
on hemostatic markers in healthy subjects or patients with coronary artery disease.
6.2.8.3. Toxicological Studies
In general, the limited toxicological studies reviewed in the 2004 PM AQCD reported positive and
negative findings for plasma fibrinogen levels or other factors involved in the coagulation cascade. Rats
exposed to New York City CAPs did not have any exposure-related effects on any measured coagulation
markers (Nadziejko et al., 2002), whereas rats exposed to a high concentration of ROFA did demonstrate
increased plasma fibrinogen (Kodavanti et al., 2002).
The coagulation effects of inhaled ultrafine carbon black (count median diameter = 36 nm) at a
concentration of 150 |ig/nr' for 6 h were evaluated 24-h post-exposure in two rat models (11-14 mo. SH
and 23 mo. Fischer 344), some of which received LPS via intraperitoneal injection prior to particle
exposure (Elder et al., 2004b). LPS has been shown to induce the expression of molecules involved in
coagulation, inflammation, oxidative stress, and the acute-phase response. In those animals only exposed
to carbon black, SH rats demonstrated increased thrombin-anti-thrombin complexes (TAT) and decreased
fibrinogen. For F344 rats, TAT complexes and fibrinogen were elevated only in those that received LPS
and carbon black. Whole-blood viscosity was not altered in either rat strain with particle exposure.
Mutlu et al. (2007) used a PM10 sample collected from Dusseldorf, Germany in mice (C57BL/6)
with and without the gene coding for IL-6. The authors report using a moderate intratracheal instillation
dose (10 (ig/mouse; roughly equivalent to 400-500 |ig/kg): the PM sample had previously been
characterized as having significant Fe, Ni, and V content (Upadhyay et al., 2003). In C57BL/6 mice, the
Dusseldorf PM shortened bleeding (32%), prothrombin (13%), and activated partial thromboplastin
(16%) times and increased platelet count, fibrinogen, and Factors II, VIII, and X activities 24-h following
exposure. The authors further demonstrated accelerated coagulation by a reduction in the left carotid
artery occlusion time (experimentally-derived by direct application of FeCl3). Additional experiments
demonstrated that IL-6" " or macrophage-depleted mice showed dramatically attenuated effects of PMi0 on
hemostatic indices, thrombin generation, and occlusion time. In IL-6"" mice, there was no change in total
cell counts or differentials in BAL fluid compared to the wild-type mice, despite the lack of IL-6. In
December 2008
6-63
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
contrast, the model of macrophage depletion had reduced levels of macrophages and IL-6 in BAL fluid,
following PM exposure. These studies suggest that instillation of Dusseldorf PM10 activates clotting
through an alveolar macrophage-dependent release of IL-6; however, other factors may also be involved
in the prothrombotic response (i.e., activation of neutrophils, other inflammatory cells, or alterations in
the levels of other cytokines).
Plasma fibrinogen levels were elevated 18 h following a single (6 h) exposure to on-road highway
aerosols when groups of rats pretreated with saline or influenza virus were combined (i.e., there was a
significant effect of particles) (Elder et al., 2004a). A PM25 CAPs exposure conducted for 2 days (4 h/day;
mean mass concentration range 144-2758 |ig/m3: August to October 2001; Research Triangle Park, NC)
in SH rats induced plasma fibrinogen increases (measured 18-20 h post-exposure) in 5 of 7 separate
studies (Kodavanti et al., 2005). Fibrinogen was not different from the air control group on the two days
with the highest CAPs concentrations (1129 and 2758 |ig/m3). indicating that the response was likely not
attributable to mass. In SH rats exposed via nose-only inhalation to PM2.5 CAPs for 6 h in one of three
locations in the Netherlands (Bilthoven-background location, Utrecht-industrial location, or a site near a
freeway tunnel and river with substantial shipping activity; mean mass concentration range 270-2400,
335-3720, and 655-3660 |ig/m3. respectively), plasma fibrinogen was increased 48-h post-exposure
(Cassee et al., 2005). A similar study conducted by the same group (Kooter et al., 2006) reported no
changes in plasma fibrinogen measured 18 h after a 2-day exposure (6 h/day) to fine or fine+ultrafine
CAPs (mean mass concentration range 399.0-1067.5 and 269.0-555.8 |ig/m3. respectively; fine CAPs site
in Bilthoven and ultrafine+fine site in freeway tunnel in Hendrik Ido Ambacht; 1/2003-4/2004).
However, elevated vWF was observed in SH rats exposed to the highest concentration of fine CAPs. In a
study employing PMi0.2.5 collected from six European locations with contrasting traffic profiles,
intratracheal instillation resulted in fibrinogen increases only in SH rats in the 10 mg/kg dose group at
24-h post-exposure; similar responses were observed in rats exposed to PM2 5 (Gerlofs-Nijland et al.,
2007).
A few studies have evaluated red blood cell (RBC) measurements following PM exposure. In
Wistar rats pre-exposed to ozone (1600 |ig/m3: 8-h) then CAPs for 6 h, increases in RBC, hemoglobin,
and hematocrit were observed 2 days after CAPs exposure (Cassee et al., 2005). For SH rats exposed to
CAPs only, decreased mean corpuscular hemoglobin concentration were reported (Cassee et al., 2005).
Decreases in mean corpuscular volume (MCV), and elevations in mean platelet volume (MPV) and mean
platelet component (MPC) were reported in SH rats 18 h following a 2-day exposure to ultrafine+fine
CAPs in a freeway tunnel (details provided above) (Kooter et al., 2006). In another study that employed
coal fly ash (mean mass concentration 1400 |ig/m3: 4 h/day><3 day) Sprague Dawley rats demonstrated
December 2008
6-64
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
increases in hematocrit and MCV in Sprague Dawley rats at 36 h but not 16-h post-exposure (Smith et al.,
2006).
Increases in coagulation and thrombotic markers were observed in some studies of rats or mice
exposed to PM. Plasma TAT complexes were increased in carbon black-exposed SH rats and shortened
bleeding, prothrombin, and activated partial thromboplastin times were observed in mice 24-h
post-exposure. Furthermore, the latter study also reported increased levels of Factors II, VIII, and X
activities in mice. Another study demonstrated increased vWF in response to PM2 5 CAPs. As for plasma
fibrinogen, these studies provide some evidence that increased plasma fibrinogen levels are observed
18 h to 48 h post-exposure to PM, although one study reported no change and another reported a decrease
in this biomarker. Alterations in platelet measurements have also been observed with PM exposure,
including increased platelet number, mean platelet volume, and mean platelet component. The
toxicological results of RBC-related measurements are limited and inconsistent following PM exposure,
which may be attributable to different exposure protocols, time of analysis, or rat strain.
6.2.9. Systemic and Cardiac Oxidative Stress
Very little information on systemic oxidative stress associated with PM was available for inclusion
in the 2004 PM AQCD. However, recent epidemiologic studies have provided consistent evidence of
PM-induced increases in markers of systemic oxidative stress including plasma TBARS, CuZn-SOD,
8-oxodG, and total homocysteine. This is supported by a limited number of human clinical studies that
observed PM-induced increases in free-radical mediated lipid peroxidation as well as upregulation of the
DNA repair gene hOGGl. In addition, recent toxicological studies have demonstrated an increase in
cardiac oxidative stress following PM exposure in rats.
6.2.9.1. Epidemiologic Studies
Although studies of markers of inflammation and coagulation were considered, no studies of
markers of oxidative stress were reviewed in the 2004 PM AQCD. Since 2002, numerous studies have
examined whether short-term increases in mean PM concentrations are associated with adverse changes
in systemic markers of oxidative stress.
In a separate analysis of the randomized trial evaluating the effect of omega-3 fatty acid
supplementation on HRV among nursing home residents living in Mexico city, Romieu et al. (2008)
investigated the effect of this intervention on markers of systemic oxidative stress (Cu/Zn SOD activity,
LPO in plasma and GSH in plasma). Supplementation with both fish oil and soy oil was related to an
increase in Cu/Zn SOD activity and GSH plasma levels. A significant decrease of Cu/Zn SOD was
December 2008
6-65
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
associated with a 10 (ig/m3 increase of PM25 in both groups [Fish oil: (3 = -0.17 (SE = 0.05), p = 0.002;
Soy oil: (3 = -0.06 (SE = 0.02) p = < 0.001], A decrease in GSH was associated with a 10 (ig/m3 increase
in PM2 5 in the fish oil group ((3 = -0.09 (SE = 0.04, p = 0.017).
Two studies evaluated plasma homocysteine levels in relation to PM (Baccarelli et al., 2007a).
Baccarelli et al. (2007a) investigated fasting and postmethionine-load total homocysteine (tHcy) among
1,213 normal subjects in Lombardia, Italy. Plasma homocysteine is a risk factor for CVD and a marker for
oxidative stress. Among smokers, average PMi0 level during the 24-h preceding the measurement was
associated with 6.3% (95% CI: 1.311.6) and 4.9% (95% CI: 0.5-9.6) increases in fasting and
postmethionine-load tHcy, respectively. No associations were observed among non-smokers. Park et al.
(2008b) investigated the association of BC, OC, sulfate and PM2 5 with tHcy among 960 male participants
of the NAS. Effect modification by folate and vitamins B6 and B12 was also examined. BC and OC were
associated with increases in tHcy and associations were more pronounced in those with lower plasma
folate and vitamin B12.
In smaller studies with 25 to 50 healthy or diseased participants, several markers of oxidative stress
have been associated with PM size fractions or components. These associations include thiobarbituric
acid reactive substances (TBARS) with 24-h PMi0 (Liu et al., 2007b); several PM metrics (e.g. ultrafine,
coarse, EC, OC, BC and PN) with Cu/Zn-SOD (Delfino et al., 2008); PM2 5, BC, vanadium and chromium
with plasma proteins (Sorensen et al., 2003); DNA damage assessed by 7-hydro-8-oxo-2-deoxyguanosine
(8-oxodG) in lymphocytes (Sorenson et al., 2005) and, 8-OHdG with sulfates (Chuang et al., 2007b).
Summary
Oxidative stress responses measured by plasma tHcy, CuZn-SOD, TBARS, 8-oxodG have been
consistently observed (Baccarelli et al., 2007a; Chuang et al., 2007b; Delfino et al., 2008; Liu et al.,
2007b; Romieu et al., 2008; Sorensen et al., 2003; Sorenson et al., 2005).
6.2.9.2. Human Clinical Studies
Brauner et al. (2007) recently investigated the effect of urban traffic particles on oxidative
stress-induced damage to DNA. Healthy adults (20-40 years old) were exposed to low concentrations of
urban traffic particles as well as filtered air for periods of 24 h, which included two 90-min periods of
exercise. Exposures took place in an exposure chamber above a busy road with high traffic density in
Copenhagen. Non-filtered air was pumped into the chamber from above the street, with average PM2 5 and
PM2 5_io mass concentrations of 9.7 |ig/m3 and 12.6 |ig/m3. respectively. The ultrafine/fine (6-700 nm)
particle number concentration was continuously monitored throughout the exposure. The PM2 5 fraction
was rich in sulfur, vanadium, chromium, iron, and copper. PBMCs were isolated from blood samples
December 2008
6-66
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
collected at 6 and 24-h. DNA damage, as measured by strand breaks (SB) and
formamidopyrimidine-DNA glycosylase (FPG) sites, was evaluated using the Comet assay. The activity
and mRNA levels of the DNA repair enzyme 7,8-dihydro-8-oxoguanine-DNA glycosylase (OGG1) were
also measured. The authors observed increased levels of DNA strand breaks and FPG sites following
6 and 24 h of exposure to PM. Using a mixed-effects regression model, the particle concentration at the
57 nm mode was found to be the major contributor of these measures of DNA damage. The results of this
study suggest that short-term (6-24 h) exposure to ambient levels of ultrafine particles cause systemic
oxidative stress resulting in damage to DNA.
In a human clinical study of controlled exposure to WS, Barregard et al. (2006) found an increase
in urinary excretion of free 8-iso-prostaglandin2a among healthy adults (n = 9) approximately 20 h
following a 4 hour exposure to PM2.5 (mass concentration of 240-280 (ig/m3). This finding provides
evidence of a PM-induced increase in free-radical mediated lipid peroxidation. From the same study,
Danielsen et al. (2008b) reported an increase in the mRNA levels of the DNA repair gene hOGGl in
peripheral mononuclear cells 20 h after exposure to WS relative to filtered air. Potential evidence of
systemic oxidative stress has also been observed following controlled human exposures to DE. Tornqvist
et al. (2007) reported an increase in plasma antioxidant capacity in a group of healthy volunteers 24-h
after a 1-h exposure to DE with a particle concentration of 300 |ig/m3. The investigators suggested that
systemic oxidative stress occurring following exposure may have caused this up-regulation in antioxidant
defense. Peretz et al. (2007) observed some significant differences in expression of genes involved in
oxidative stress pathways between exposure to DE (200 (.ig/nr3 PM2 5) and filtered air. However, the
conclusions of this investigation are limited by a small number of subjects.
Based on the results of these studies, it appears that exposure to PM at or near ambient levels may
increase systemic oxidative stress in human subjects.
6.2.9.3. Toxicological Studies
Very little information was available for inclusion in the 2004 PM AQCD on oxidative stress in the
cardiovascular system. A few new studies have evaluated ROS in blood or the heart following PM
exposure. Some studies have used chemiluminescence (CL), which is measured using the decay of
excited states of molecular oxygen, and may also be prone to artifact.
Gurgueira et al. (2002) measured oxidative stress in Sprague Dawley rats immediately following a
5-h CAPs exposure (PM2 5; mean mass concentration range 99.6-957.5 |ig/m3: Boston, MA; July 2000 to
February 2001) and reported increased in situ CL in hearts of CAPs-exposed animals. CL evaluated after
CAPs exposure durations of 1 and 3 h did not demonstrate changes from the filtered air group, although a
5-h exposure resulted in increased CL in hearts. When animals were allowed to recover for 24-h prior to
December 2008
6-67
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
CL evaluation, oxidative stress returned to control values. To compare potential particle-induced
differences in in situ CL, rats were exposed to ROFA (1.7 mg/m3 for 30 min) or carbon black (170 |ig/m3
for 5 h) and only the ROFA-treated animals exhibited increased CL in cardiac tissue. Additionally, levels
of antioxidant enzymes in the heart (Cu/Zn superoxide dismutase (SOD) and MnSOD) were increased
(100% and -40%, respectively) in CAPs-exposed rats.
Recently, Rhoden et al. (2005) tested the role of the ANS in driving CAPs-induced cardiac
oxidative stress in heart tissues of Sprague Dawley rats. At mass concentrations of 700 (ig/m3 (Boston,
MA), pretreatment with N-acetylcysteine (an antioxidant; NAC), atenolol (a Pi-receptor antagonist), or
glycopyrrolate (a muscarinic receptor antagonist) attenuated the CL and TBARS effects observed in the
heart following a 5-h PM2 5 exposure. The wet/dry ratio (edema) of cardiac tissue also returned to control
values in animals treated with NAC prior to CAPs. These combined results indicate involvement of both
the sympathetic and parasympathetic pathways in the cardiac oxidative stress response observed
following PM exposure.
More recently, a type of irritant receptor, the Transient Receptor Potential Vanilloid Receptor 1
(TRPV1), was identified as central to the inhaled CAPS-mediated induction of cardiac tissue CL and
TBARS (Ghelfi et al., 2008). In these studies (Sprague Dawley rats; mean CAPs concentration
218 (.ig/nr3; mean CAPs concentration range 100-550 (ig/m3; Boston, MA), capsazapine (aTRPVl
inhibitor) abrogated cardiac CL, TBARS, edema, and QT-interval shortening when measured at the end of
the 5-h exposure. These studies provide some evidence that the ANS may be involved in producing
cardiac oxidative stress following exposure to CAPs. Furthermore, this response could be acting, at least
in part, via TRPV receptors.
In Wistar Kyoto rats exposed to CAPs in Japan, relative mRNA expression of HO-1 was increased
in cardiac tissue and was also significantly correlated with the cumulative mass of PM collected on
chamber filters throughout the exposure (Ito et al., 2008). Other studies presented in earlier sections also
demonstrated ROS (via CL) and nitrotyrosine expression (via ELISA) in the left ventricle with carbon
black exposure (Tankersley et al., 2008) and oxidative stress in the systemic microcirculation (via
tetranitroblue tetrazolium reduction method) following ROFA intratracheal instillation exposure
(Nurkiewicz et al., 2006).
When considered together, the above studies provide some evidence that PM exposure results in
oxidative stress as measured in cardiac tissue by CL, TBARS, HO-1 mRNA expression, and nitrotyrosine
expression. Cardiac oxidative stress may have resulted from PM stimulation of the ANS, although these
studies have only been conducted in one laboratory. A single study provided support for vascular
oxidative stress as demonstrated in the microcirculation following ROFA exposure.
December 2008
6-68
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
PM Components
Individual PM component concentrations were linked to CL levels in rat heart tissue using separate
univariate linear regression models, with total PM mass, Al, Si, Ti, and Fe having p-values <0.007
(Gurgueira et al., 2002). The highest R2 value in the regression analyses was for Al (0.67) and its
concentration ranged from 0.000 to 8.938 (ig/m3.
6.2.10. Hospital Admissions and ED Visits
The 1996 PM AQCD (1996) considered just two time-series studies regarding the association
between daily variations in PM levels and the risk of cardiovascular disease (CVD) morbidity as
measured by the number of daily hospitalizations with primary discharge diagnoses related to CVD
(Burnett et al., 1995; Schwartz and Morris, 1995). In contrast, the 2004 PM AQCD (U.S. EPA, 2004)
reviewed more than 25 publications relating PM and risk of CVD hospitalizations. Results from a handful
of larger multicity studies were emphasized, with the greatest emphasis placed on findings from the U.S.
National Morbidity, Mortality, and Air Pollution Study (NMMAPS) (Samet et al., 2000) and a subsequent
reanalysis (Zanobetti and Schwartz, 2003). The NMMAPS study evaluated the effect of daily changes in
ambient PM levels on total CVD hospitalizations among elderly Medicare beneficiaries in 14 U.S. cities
and found a -1% excess risk per 10 (ig/m3 increase in PM10. The 2004 PM AQCD concluded that these
results, along with those of the other single- and multicity studies reviewed "generally appear to confirm
likely excess risk of CVD-related hospital admissions for U.S. cities in the range of [0.6 to 1.7% per
10 |_ig/m31 PMio, especially among the elderly" (U.S. EPA, 2004). The 2004 PM AQCD also concluded
that there was some evidence from single city studies suggesting an excess risk specifically for
hospitalizations related to ischemic heart disease and heart failure. Furthermore, the 2004 PM AQCD
found that "insufficient data exist from the time-series CVD admissions studies [... ] to provide clear
guidance as to which ambient PM components, defined on the basis of size or composition, determine
ambient PM CVD effect potency" (U.S. EPA, 2004). The key studies reviewed in the 2004 PM AQCD on
this topic included those by Burnett and colleagues (1997; 1999), Lippman and colleagues (2000), Ito
(2003), and Peters et al. (2001).
Recent large studies conducted in the U.S., Europe, and Australia and New Zealand have confirmed
these findings for PMi0, and have also observed consistent associations between PM2 5 and cardiovascular
hospitalizations. However, findings from single city studies have demonstrated regional heterogeneity in
effect estimates. It is apparent from these recent studies that the observed increases in cardiovascular
hospitalizations are largely due to admissions for ischemic heart disease and congestive heart failure
rather than cerebrovascular diseases. The new literature on hospitalizations and ED visits for
December 2008
6-69
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
cardiovascular causes published since 2002 is reviewed in the following sections. First, the specific CVD
outcomes captured using ICD codes from hospital admissions databases are discussed. Second, the
methods used in the large and multicity studies are described. The remainder of the chapter is organized
as follows: (1) evidence PM effects by outcome (e.g. all CVD, cardiac diseases, IHD, MI, CHF,
arrhythmias, cerebrovascular diseases, ischemic and hemorrhagic stroke, peripheral vascular disease); (2)
evidence of PM effects in susceptible populations; and (3) overall summary. Within each section,
evidence from large/multi-city studies is emphasized and results from U.S. and Canadian single city
studies are also discussed. Although the single city studies may lack statistical power needed to evaluate
interactions and detect some of the subtle effects of air pollution, they inform our interpretation of the
heterogeneous effect estimates that have been observed across North America.
Cardiovascular Disease ICD Codes
When the 2004 PM AQCD was written, few studies had evaluated the link between ambient PM
and specific CVD outcomes such as congestive heart failure, ischemic heart disease or ischemic stroke. In
contrast, the majority of recent studies have focused on specific CVD outcomes. This trend is justified by
the fact that the short-term exposure effects of PM may be very different for different cardiovascular
outcomes. For example, given the current putative biological pathways involved in the acute response to
PM exposure, there is no a priori reason why short-term fluctuations in PM levels would have similar
effects on the risk of acute MI, chronic atherosclerosis of the coronary arteries, and hemorrhagic stroke.
Almost all of the published time-series studies of cardiovascular hospitalizations and emergency
department visits identified cases based on administrative discharge diagnosis codes as defined by the
International Classification of Disease 9th revision (ICD-9) or 10th revision (ICD-10) (NCHS, 2007). A
complicating factor in interpreting the results of these studies is the lack of consistency in both defining
specific health outcomes and in the nomenclature used.
Table 6-5 shows major groups of diagnostic codes used in air pollution studies for diseases of the
circulatory system. The codes ICD-9: 390-459 are frequently used to identify all CVD morbidity. Note
that this definition of CVD includes diseases of the heart and coronary circulation, cerebrovascular
disease, and peripheral vascular disease. In contrast, the term cardiac disease specifically excludes
diseases not involving the heart or coronary circulation. While this distinction is conceptually
straightforward, the implementation of the definition of cardiac disease in terms of ICD-9 or ICD-10
codes varies among authors. Even greater heterogeneity can be found among studies in the
implementation of definitions related to cerebrovascular disease.
December 2008
6-70
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Table 6-5. Description of ICD-9 and ICD-10 codes for diseases of the circulatory system.
Description
ICD-9 Codes
ICD-10 Codes
All Cardiovascular Disease
390-459
IOO-I99
Ischemic Heart Disease
410-414
I20-I25
Acute Myocardial Infarction
410
121
Diseases Of Pulmonary Circulation
415-417
I26-I28
Heart Failure
428
I50
Arrhythmia
427
I47,148,149
Cerebrovascular Disease
430-438
I60-I69
Ischemic Stroke And Transient Ischemic Attack (TIA)
430-432
I63
Hemorrhagic Stroke
433-435
I60-I62
Peripheral Vascular Disease (PVD)
440-448
I70-I79
Design and Methods of Large and Multicity Hospital Admission and Emergency
Department Visit Studies
Recently, multiple research groups in the U.S., Europe, and Australia have created large datasets to
evaluate specific CVD and respiratory endpoints using more detailed and relevant measures of PM
concentration. In the U.S., the MCAPS analyses of Dominici et al. (2006), Bell et al. (2008a) and Peng et
al. (2008) are large, comprehensive and informative studies based on Medicare hospitalization data.
Likewise, the Atlanta-based SOPHIA study (Metzger et al., 2004; Peel et al., 2005) is the largest and most
comprehensive study of U.S. cardiovascular and respiratory ED visits. In Europe, the APHEA initiative
(Le Tertre et al., 2002a; 2003), the more recent HEAPSS study (von Klot et al., 2005), and the French
PSAS program (Host et al., 2008; Larrieu et al., 2007) are similarly noteworthy for their large sample
size, geographic diversity, and consideration of specific CVD and/or respiratory endpoints. These studies
contain adequate data to examine interactions by season and region; the effects of different size fractions,
components and sources of PM; or the effect of PM on susceptible subpopulations. The following section
provides a detailed review of the study design and methods used by each of the large studies. A discussion
of the results of each study can be found in later sections of the ISA.
MCAPS: Medicare Air Pollution Study
Dominici et al. (2006) created a database of daily time-series for 1999 through 2002 of hospital
admission rates for a range of cardiovascular and respiratory outcomes among Medicare beneficiaries
aged > 65 years, ambient PM2 5 levels, and meteorological variables for 204 U.S. urban counties. The
specific CVD outcomes considered were: cerebrovascular disease (ICD-9: 430-438), peripheral vascular
December 2008
6-71
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
disease (440-448), ischemic heart diseases (410-414, 429), heart rhythm disturbances (426, 427), and
heart failure (428). Injuries (800-849) were evaluated as a control outcome. Gaseous and other particulate
pollutant size fractions were not considered.
Data on PM25 were obtained from the AQS database of the U.S. EPA. Within each county,
associations between cause-specific hospitalization rates and same-day PM2.5 levels were evaluated using
Poisson regression models controlling for long-term temporal trends and meteorologic conditions with
natural cubic splines. County-specific results were subsequently averaged using Bayesian hierarchical
models. In addition to evaluating single day lags, three-day distributed lag models (lags 0, 1, and 2 days)
were also considered in a subset of 90 U.S. counties with daily PM2 5 data available during the study time
period.
Subsequently, Peng et al. (2008) and Bell et al. (2008a) extended the database of daily time-series
of hospital admissions, PM2 5, and other covariates for 202 U.S. counties through 2005. Importantly, Peng
et al. (2008) added data on PM10-2.5 to this database for 108 U.S. counties with one or more co-located
PM2 5 and PMi0 monitors. Analyses with PM10-2.5 were carried out using similar methods to those of
Dominici et al. (2006). Peng et al. (2008) evaluated the robustness of PM2 5 associations to adjustment for
thoracic coarse PM (Peng et al., 2008). Gaseous pollutants were not considered in these analyses.
SOPHIA: Study of Particulates and Health in Atlanta
SOPHIA investigators (Metzger et al., 2004; Peel et al., 2005; 2007; Tolbert et al., 2000) compiled
data on 4,407,535 emergency department (ED) visits between 1993 and 2000 to 31 hospitals in the
Atlanta metropolitan statistical area (20 counties). Specific cardiovascular outcomes considered were:
ischemic heart disease (ICD-9: 410-414), acute myocardial infarction (410), cardiac dysrhythmias (427),
cardiac arrest (427.5), congestive heart failure (428), peripheral vascular and cerebrovascular disease
(433-437, 440, 443-444, 451-453), atherosclerosis (440), and stroke (436). Finger wounds (883.0) were
evaluated as a control outcome.
The air quality data included measurements of criteria pollutants (PM and gaseous pollutants) for
the entire study period, as well as detailed measurements of mass concentrations for the fine (PM2 5) and
thoracic coarse fractions (PM10-2.5) of PM and several physical and chemical characteristics of PM2 5 for
the final 25 months of the study using data from the ARIES monitoring station. Rates of ED visits for
specific causes were assessed in relation to the 3-day moving average (lags 0-2 days) of daily measures of
air pollutants using Poisson generalized linear models controlling for long-term temporal trends and
meteorologic conditions with cubic splines. Tolbert et al. (2007) published interim results of this study in
relation to both cardiovascular and respiratory disease visits, Metzger et al. (2004) published the main
results for CVD visits, and Peel et al. (2005) published the main results for respiratory conditions. An
December 2008
6-72
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
analysis of co-morbid conditions that may make individuals more susceptible to PM-related
cardiovascular risk was carried out by Peel et al. (2007). Tolbert et al. (2007) extended the available data
through 2002 and compared results from single and multipollutant models while Sarnat et al. (2008)
evaluated the risk of ED visits for cardiovascular and respiratory diseases in relation to specific sources of
ambient PM using the extended dataset.
APHEA and APHEA-2: Air pollution and Health: a European Approach
APHEA-2 investigators compiled daily data on cardiovascular (Le Tertre et al., 2002b, 2003) and
respiratory (Atkinson et al., 2001; 2004) disease hospital admissions in the following 8 European cities:
Barcelona, Birmingham, London, Milan, the Netherlands, Paris, Rome, and Stockholm. (The publications
on respiratory diseases were reviewed in the 2004 PM AQCD.) The specific CVD outcomes considered in
each city were: cardiac diseases (ICD-9: 390-429), ischemic heart disease (410-413) and cerebrovascular
diseases (430-438). Routine registers in all cities provided daily data on hospitalizations. Only emergency
hospitalizations were considered, except in Milan, Paris, and Rome where only general admissions data
were available.
Ambient PM10 levels were available in all cities except Paris (PM13 used), and Milan and Rome
(total suspended particulates [TSP] used). Data on gaseous pollutants (N02, S02, CO, and 03) were also
available in most cities. Five of the eight cities provided data on black smoke (BS). The length of the
available time-series varied by city but generally spanned from the early to mid 1990s.
Within each city, associations between cause-specific hospitalization rates and same-day PM2 5
levels were evaluated using Poisson generalized additive models (GAMs) controlling for long-term
temporal trends and meteorologic conditions. City-specific results were subsequently averaged using
standard meta-analytic methods. The original analyses (Atkinson et al., 2001; Le Tertre et al., 2002a)
were carried out using GAMs and LOESS smoothers. Following reports of problems associated with
using the default convergence criteria in the standard S-plus GAM procedure (Dominici et al., 2002),
study authors reanalyzed the data on cardiac admissions using GAMs and stricter convergence criteria,
and generalized linear models (GLMs) with natural splines and penalized splines (Atkinson, 2004; Le
Tertre et al., 2003). The authors found that the results of the original analyses were insensitive to the
choice of convergence criteria and that the use of GLMs with penalized splines yielded very similar
results.
HEAPSS: Health Effects of Air Pollution among Susceptible Subpopulations
HEAPSS investigators collected data on patients hospitalized for a first myocardial infarction (MI)
in five European cities between 1992 and 2000. Patients were identified from MI registers in Augsburg
December 2008
6-73
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
and Barcelona, and from hospital discharge registers in Helsinki, Rome and Stockholm. Data on daily
levels of PM10, were measured at central monitoring sites in each city. Particle number concentration, a
proxy for ultrafine particles, was measured for a year in each city and then modeled retrospectively for the
whole study period. Associations of outcomes with gaseous criteria pollutants were also evaluated.
Von Klot et al. (2005) identified 22,006 survivors of a first MI in the five participating European
cities and collected data on subsequent first cardiac re-hospitalizations between 1992 and 2001.
Readmissions of interest were those with primary diagnoses of acute MI, angina pectoris, or cardiac
disease (which additionally includes dysrhythmias and heart failure). Within each city, associations
between cause-specific hospitalization rates and same-day levels of PMi0 were evaluated using Poisson
GAMs controlling for long-term temporal trends and meteorologic conditions using penalized splines.
City-specific results were combined using standard meta-analytic methods. Subsequently, Lanki et al.
(2006a) used HEAPSS data from 26,854 patients to evaluate the association between daily PMi0 and
particle number concentrations and the risk of hospitalization for first MI.
PSAS: The French National Program on Air Pollution Health Effects
Larrieu et al. (2007) evaluated the association between PM10 and the risk of hospitalization in 8
French cities between 1998 and 2003. The cities examined were: Bordeaux, Le Havre, Lille, Lyon,
Marseille, Paris, Rouen and Toulouse. The specific CVD outcomes considered in each city included: total
CVD (ICD-10: 100-199), cardiac disease (100-152), ischemic heart diseases (120-125) and stroke (160-164,
G45-G46). The available data did not differentiate between emergency and non-emergency
hospitalizations. Daily mean PMi0 and N02 levels as well as 8-h maximum ozone levels were obtained
from a network of monitors in each city.
Within each city, associations between cause-specific hospitalization rates and 2-day moving
average (lag 0-1 days) levels of PMi0 were evaluated using Poisson GAMs controlling for long-term
temporal trends and meteorologic conditions using penalized splines. City-specific results were combined
using standard meta-analytic methods. Host et al. (2008) used a subset of these data (6 cities, 2000-2003)
to compare the effects of the fine (PM2 5) and course fractions (PMi 0-2.5) of ambient particles on the risk of
cardiovascular and respiratory admissions. CVD outcomes assessed in this analysis were all CVD
(ICD-10 100-199), cardiac (100-152) and IHD (120-125). PM2 5 levels were obtained from the same
network of background monitors described above. PM10.2.5 was calculated by subtracting PM2 5 levels
from PM10 levels. Gaseous pollutants and hospital admissions for stroke were not considered in this
analysis.
December 2008
6-74
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Multicity Studies in Australia and New Zealand
Barnett et al. (2006) collected data on daily CVD emergency hospital admissions and pollution data
between 1998 and 2001 in five Australian cities (Brisbane, Canberra, Melbourne, Perth, Sydney) and two
cities in New Zealand (Auckland, Christchurch). In 2001, these cities covered 53% of the Australian
population and 44% of the New Zealand population. The specific outcomes considered in each city were:
all circulatory diseases (ICD-9 390-429, ICD-10 100-199 with exclusions); heart failure (ICD-9 428,
ICD-10 150); arrhythmia (ICD-9 427 ICD-10 146-49); cardiac disease (ICD-9 390-429, ICD-10 100-152,
197.0,197.1,198.1); ischemic heart disease (ICD-9 410-413, ICD-10 120-24,125.2); acute MI (ICD-9 410,
ICD-10 121-22); and stroke (ICD-9 430-438, ICD-10 160-66,167,168,169, G45-46 with exclusions).
Air pollutants considered were 24-h average PMi0, 24-h average PM25, BSP and gaseous
pollutants. Within each city, associations between cause-specific hospitalization rates and 2-day moving
average (lags 0-1 days) of PM10 were evaluated using the time-stratified case-crossover approach which
controls for long-term and seasonal time trends by design rather than analytically. City-specific results
were combined using random effects meta-analytic methods.
EMECAS: Spanish Multicentric Study on the Relation between Air Pollution and Health
Ballester et al. (2006) collected data on daily cardiovascular emergency hospital admission and air
pollution data between approximately 1995 and 1999 in 14 cities in Spain. The specific outcomes
considered in each city were: total CVD (ICD-9: 390-459) and heart diseases (410-414, 427, 428). Air
pollutants considered were PMi0, TSP, BS, S02, N02 (24-h averages), CO and ozone (8-h maximums).
Within each city, associations between cause-specific hospitalization rates and daily levels of each
pollutant metric were evaluated using Poisson GAMs with strict convergence criteria. In all models,
pollutants were entered as linear continuous variables and included control for confounding by
meteorological variables, influenza rates, long-term time trends, and unusual events. The authors
considered both distributed lag models (lags 0-3 days) and the 2-day moving average of pollution (lags
0-1 days). City-specific results were combined using standard meta-analytic methods.
6.2.10.1. All Cardiovascular Disease
The 2004 PM AQCD incorporated the results of a large number of time-series studies in the U.S.
and elsewhere relating ambient PM levels and risk of hospitalization for CVD. The 2004 PM AQCD
noted that the strongest evidence for this association came from the NMMAPS study (Samet et al., 2000)
and the subsequent reanalysis by Zanobetti and Schwartz (2003).
Since then, the U.S. MCAPS study evaluated the association between PM2 5 and risk of CVD
hospitalization in 202 U.S. counties between 1999 and 2005 and found a 0.7% (95% posterior interval
December 2008
6-75
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
(PI): 0.5, 1.0) increase in risk per 10 (ig/m3 increase in PM25 on the same day (Peng et al., 2008). In 108
U.S. counties with co-located PM10 and PM10.2.5 monitors, the authors found a 0.4% (95% PI, 0.1 to 0.7,
lag 0) increase in risk per 10 |_ig/nr' PM10.2.5 (Peng et al., 2008). In a 2-pollutant model adjusted for PM25,
the association between PM10-2 5 and CVD hospitalization lost precision (0.3% [95% PI: -0.1 to 0.6])
(Peng et al., 2008). Bell et al. (2008a) found evidence of substantial and statistically significant variability
in the effects of PM2 5 on cardiovascular hospitalizations by season and region, with the highest national
average estimates occurring in the winter and the highest regional estimates in the Northeastern U.S.
(1.08% 95%PI: 0.79, 1.37, lag 0, per 10 (ig/m3 increase in PM25). Estimates for the nation (1.49%
(95%PI: 1.09, 1.89, lag 0) and northeast (2.01% 95%PI 1.39, 2.63, lag 0) were highest in the winter.
Additional evidence is provided by several large multicity studies conducted outside of the U.S.
The European APHEA2 study (Le Tertre et al., 2002a) looked at admissions for CVD (defined as ICD-9
390-429) among those aged > 65 and found a 0.7% (95% CI: 0.4, 1.0, lag 0-1 day average) increase in
risk per 10 (ig/m3 PMi0. The Spanish EMECAS study (Ballester et al., 2006) looked at admissions for
CVD (defined as ICD-9 390-459) and found a 0.9% (95% CI: 0.4, 1.5, lag 0-1 day average) increase in
risk per 10 (ig/m3 PMi0. The French PSAS program looked at CVD hospitalizations (defined as ICD-10
100-199) among the elderly and found a 1.1% (0.5, 1.7%) increase in risk with PMi0 and a 1.9% (95% CI:
0.9, 3.0, lag 0-1 day average) increase in risk with a 10 (ig/m3 increase in PM25 (Host et al., 2007; Larrieu
et al., 2007). Non-significant increases in association with PM10.2 5 were reported (1.0% [95% CI: -1.0 to
3.0]) (Host et al., 2008). In multiple cities across New Zealand and Australia, Barnett et al. (2006) looked
among the elderly (CVD defined as ICD-9 390-459 and found a 1.3% (95% CI: 0.6-2.0, lag 0-1 day
average) increase in risk per 10 (ig/m3 increase in PM2 5.
The Atlanta-based SOPHIA study found a 0.9% (95% CI: -0.2 to 1.9, lag 0-2 d average) and a 3.3%
(95% CI: 1.0-5.6, lag 0-2 d average) increase in risk with a 10 (ig/m3 increase in PMi0 and PM25,
respectively (Metzger et al., 2004). In a more recent analysis from this study with an additional 4 years of
data, ED visits for CVD were not significantly associated with PMi0 or PM2 5, but were significantly
associated with total carbon (1.6% [95% CI: 0.5-2.6, per IQR increase]), EC (1.5% [95% CI: 0.5-2.5, per
IQR increase]) and organic carbon (1.5% [95% CI: 0.5-2.6, per IQR increase]) components of PM2 5
(Tolbert et al., 2007). More recently, Sarnat et al. (2008) used multiple source-apportionment methods to
evaluate the association between all CVD ED visits and specific PM2 5 sources and found consistent
positive associations with sources related to motor vehicles and biomass combustion. These results were
insensitive to the source-apportionment technique used. It is noteworthy that other traffic-related gaseous
pollutants were associated with CVD ED visits in the SOPHIA study (Metzger et al., 2004).
Using meta-regression techniques and the reported association between PMi0 and CVD
hospitalizations from the 14 cities included in the NMMAPS analysis, Janssen et al. (2002) examined
December 2008
6-76
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
whether the between-city variability in relative risk estimates were related to the local contribution of a
number of PM sources. The authors found that in multivariate analyses PM10 coefficients increased
significantly with increasing percentage of PM10 emissions from highway vehicles/diesels and oil
combustion.
A small number of additional single-city studies have been published showing positive associations
between hospital admissions and ambient PM in Copenhagen, Denmark (Andersen et al., 2007b), and
weak nonsignificant associations in Spokane, WA (Schreuder et al., 2006; Slaughter et al., 2005) and two
small counties in Idaho (Ulirsch et al., 2007). Schreuder et al. (2006) performed a source apportionment
analysis using seven years of daily speciation data from the same residential monitor in Spokane, WA
used by Slaughter et al. (2005). These authors related daily levels of four sources (woodsmoke, an As-rich
source, motor vehicle emissions, and airborne soil) to the excess risk of cardiovascular ED visits. During
the heating season the only notable association for CVD-related ED visits was with woodsmoke, while in
the non-heating season the only notable association was with airborne soil. While neither of these
associations reached statistical significance, the study likely lacked the statistical power to find effects of
the expected magnitude. In fact, it is doubtful that studies conducted outside of large metropolitan areas
have sufficient statistical power to detect associations of the expected magnitude.
Studies in several cities in Australia have investigated the association of cardiovascular disease
admissions with PM concentration and sources. A study from Sydney, Australia found a 0.3% (95% CI:
-0.8 to 1.4) and 1. 8% (95% CI: 0.4-3.2) excess risk per 10 (ig/m3 increase in the 2-day moving average
(lags 0-1 days) in PM10 and PM2 5, respectively (Jalaludin et al., 2006) Johnston et al. (2007) and Hanigan
et al. (2008) studied the association between PM10 and cardiovascular and respiratory hospitalizations in
Darwin, Australia, where the predominant source of PM is from biomass combustion. The authors found
little or no evidence of an association between PMi0 and cardiovascular disease hospital admissions in the
general population.
Crustal material has also been investigated in an effort to explain associations of PM concentration
with cardiovascular disease admissions. Studies of a dust storm in the Gobi desert that transported PM
across the Pacific Ocean reaching the western U.S. in the spring of 1998 have been conducted. An
analysis of the health impacts of this event on the population of British Columbia's (Canada) Lower
Fraser Valley found no excess risk of cardiac or respiratory hospital admissions despite hourly PMi0
levels >100 (ig/m3 (Bennett et al., 2006). On the other hand, a number of studies in Asia and Eastern
Europe have identified some adverse health effects associated with dust storm events. Middleton et al.
(2008) found that dust storms in Cyprus were associated with a 4.7% (95% CI: 0.7-9.0) and 10.4% (95%
CI: -4.7 to 27.9) increase in risk of hospitalization for all causes and cardiovascular diseases, respectively.
Chan et al. (2008) studied the effects of Asian dust storms on cardiovascular hospital admissions in
December 2008
6-77
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Taipei, Taiwan and also found significant adverse effects during 39 Asian dust events with high PM10
levels (daily PM10 >90 |_ig/m3). Bell et al. (2008) analyzed these data independently and concluded that
Asian dust storms were positively associated with risk of hospitalization for ischemic heart disease.
The effect estimates from multicity studies and single city studies conducted in the U.S. and
Canada are included in Figure 6-1. Information on PM concentrations during the relevant study period is
presented in Table 6-6. In summary, large studies from the U.S., Europe, and Australia/New Zealand
published since the 2004 PM AQCD provide support for an association between short-term increases in
ambient levels of PMi0 and PM2 5 and increased risk of hospitalization for total CVD. The average excess
risk among the U.S. elderly is likely in the range of 0.5 to 1.0% per 10 (ig/m3 increase in PM2 5, although
substantial variability by region of the country and season has been demonstrated. An excess risk of ED
visits for CVD of a similar magnitude appears likely. The excess risk of CVD hospitalization may be
somewhat greater in Europe and Australia/New Zealand than in the U.S. Sources including wood burning,
oil burning, traffic and crustal material have been associated with increases in cardiovascular
hospitalization or ED visits.
December 2008
6-78
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Location
Lag
Zanobetti & Schwartz 2003
10 US Cities
0-1
Metzger et al 2004* (All Ages)
Atlanta, GA
0-2
Tdbert et al 2007* (All Ages)
Atlanta, GA
0-2
Ba Hester et al 2006
14 Cities Spain
0-1
Ulirscli et al 2007 (50+ yrs)
2 Cities Idaho
0
Slaughter et al 2005 (All ages)
Spokane, WA
1
Le Tertre 2002
8 Cities Europe
0-1
Bell et al. 2008
202 US Counties
0
0
n
Host et al 2008
6 French Cities
u
0
0-1
Barnett etal 2006
7 Cities Australia NZ
0-1
Metzger et al 2004* (All ages)
Atlanta, GA
0-2
Tolbert et al 2007* (All Ages)
Atlanta, GA
0-2
Slaughter et al 2005 (All ages)
Spokane, WA
1
Peng et al. 2008
Host et al 2008
Tdbert et al 2007* (All Ages)
108 US Counties
6 Cities France
Atlanta, GA
0
1
2
0-1
0-2
PM10
PM2l;
PM
10-2.5
¦ Overall
»_NE
» Winter
» Winter, NE
-1—i—i—i—I—i—i—i—r~
-5 -3 -1 0 1 2 3 4 5
Excess Risk Estimate
Figure 6-1. Excess risk estimates per 10 pg/m3 increase in PM10, PM2.5 and PM10 2.5 for studies of
CVD ED visits * and hospitalizations. Studies represented in the figure include all
multicity studies. Single city studies conducted in the U.S. or Canada are also
included.
December 2008
6-79
DRAFT-DO NOT QUOTE OR CITE
-------
Table 6-6. Characterization of ambient PM concentrations in studies of hospital admission and
ED visits for cardiovascular diseases.
Pollutant Author
Location
Mean Concentration
(|jg/m3)
Upper Percentile Concentration
(|jg/m3)
PMw
Ballesteret al. (2006)
14 cities in Spain
NR
NR
Barnett et al. (2006)
5 cities in Australia and New
Zealand
NR
NR
Burnett et al. (1999)
Toronto, Canada
30.2
Ito et al. (2003); Lippmann
(2000)
Detroit, Ml
31
Jalaludin et al. (2006)
Sydney, Australia
16.8
75th: 19.9
Max: 103.9
Larrieu et al. (2007)
8 cities in France
Le Tertre et al. (2002a)
8 cities in Europe
Range: 15.5-55.7
Range 75th: 19.9-66
Linn et al. (2000)
Los Angeles, California
45
78 (summer) -132 (autumn)
Metzger et al. (2004)
Atlanta, GA
26.3
90th: 44.7
Morris et al. (1998)
Chicago, Illinois
41
75th: 51
Max: 117
Peters et al. (2001)
Boston, MA
Schwartz et al. (1995)
Detroit, Ml
48
90th: 82
Slaughter etal. (2005)
Spokane, WA
NR
90th: 41.9
Tolbert et al. (2007)
Atlanta, GA
Ulirsch et al. (2007)*
2 cities in southeast Idaho
24.2/23.2
90th: 40.7/37.4
Wellenius et al. (2005b)
Pittsburgh, PA
31.1
95th: 70.5
Wellenius et al. (2006b)
7 cities in the U.S.
28.3 (median)
90th: 57
Zanobetti and Schwartz
(2005)
Boston, MA
28.4 (median)
90th: 53.6
Wellenius et al. (2005b)
9 cities in the U.S.
28.4 (median)
90th: 57.9
PMis
Belief al. (2008a)
202 counties in the U.S.
NR
NR
Host et al. (2008)
6 cities in France
13.8-18.8
NR
Barnett et al. (2006)
7 cities in Australia
Metzger et al. (2004)
Atlanta, GA
17.8
90th: 32.3
Tolbert et al. (2007)
Atlanta, GA
17.1
Slaughter etal. (2005)
Spokane, WA
NR
90th: 20.2
Dominici et al. (2006)
204counties in the U.S.
NR
NR
Burnett et al. (1999)
Toronto Canada
18
Ito et al. (2003); Lippmann
(2000)
Detroit, Ml
18
Pope et al. (2006)
Wasatch Front, Utah
Symons et al. (2006)
Baltimore, MD
16
69.2
Villenueve et al. (2006)
Edmonton, Canada
8.5
75th: 11
PMlO-25
Burnett et al. (1999)
Toronto, Canada
12.2
Peng et al. (2008)
204 cities in the U.S.
NR
NR
December 2008
6-80
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Le Tertre et al. (2002a) 8 cities in Europe NR NR
Tolbert et al. (2007) Atlanta, GA 9
Host et al. (2008) 6 cities in France 7-11 NR
Peters et a I. (2001) Boston, MA 7.4
Ito et al. (2003); Lippmann Detroit, Ml 13
(2000)
Metzger et al. (2004) Atlanta, GA 9.1 90th: 16.2
Slaughter etal. (2005) Spokane, WA NR NR
* Results presented separately for 2 separate time series
6.2.10.2. Cardiac Diseases
Cardiac disease represents a subset of CVD which specifically excludes hospitalizations for
cerebrovascular disease, peripheral vascular disease, and other circulatory diseases not involving the heart
or coronary circulation. Only a small number of studies published since the 2004 PM AQCD have
evaluated the association between ambient PM and hospitalizations for cardiac diseases, as most
investigators have focused instead on more narrowly defined outcomes.
The French PSAS program found a 1.5% (95% CI: 0.5, 2.2, lag 0-2) and 2.4% (95% CI: 1.2, 3.7,
lag 0-2) excess risk among the elderly per 10 (ig/m3 increase in PMi0 and PM2 5, respectively (Host et al.,
2007). The European HEAPS S study looked at cardiac readmissions among survivors of a first MI and
found a 2.1% (95% CI: 0.4, 3.9, lag 0) excess risk per 10 (ig/m3 increase in PMi0 (von Klot et al., 2005).
A 1.9% (95% CI: 1.0, 2.7, lag 0-1) excess risk per 10 (ig/m3 increase in PM2.5 was observed in several
cities in Australia and New Zealand (Barnett et al., 2006). Single-city studies of hospital admissions from
Kaohsiung and Taipei, Taiwan, and an ED visit study from Sydney, Australia also reported statistically
significant positive associations (Chang et al., 2005c; Jalaludin et al., 2006; Yang et al., 2004a).
In summary, large studies from Europe and Australia/New Zealand published since the 2004 PM
AQCD provide support for an association between short-term increases in ambient levels of PMi0 and
PM2 5 and increased risk of hospitalization for cardiac disease. The excess risk for cardiac hospitalizations
may be somewhat larger than for total CVD hospitalizations.
6.2.10.3. Ischemic Heart Disease
IHD represents a subset of all cardiac disease hospitalizations and typically includes acute MI (ICD
9: 410), other acute and subacute forms of ischemic heart disease (411), old MI (412), angina pectoris
(413), and other forms of chronic ischemic heart disease (414). Some authors term this category coronary
heart disease. Published studies evaluating IHD as a single outcome are considered first, followed by
consideration of studies looking at acute MI, a specific form of IHD.
December 2008
6-81
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
In 1995 Schwartz and Morris (1995) reported a 0.6% (95% CI: 0.2, 1.0%) excess risk of
hospitalization for IHD per 10 (ig/m3 increase in mean PM10 levels over the past two days among elderly
Medicare beneficiaries living in Detroit between 1986 and 1989. As reviewed in the 2004 PM AQCD,
similar associations were subsequently observed in many single-city studies including: London, England
(Atkinson et al., 1999), Toronto, Canada (Burnett et al., 1999), and Seoul, Korea (Lee et al., 2003).
Studies in Hong Kong (Wong et al., 1999; 2002b), Birmingham, England (Anderson et al., 2001), and
London, England (Wong et al., 2002a) yielded positive point estimates of a similar magnitude, but did not
reached statistical significance.
The positive associations between short-term changes in PM and IHD hospitalizations observed in
the early single-city studies have been confirmed in several large multicity studies. The U.S. MCAPS
study (Dominici et al., 2006) found a 0.4% (95% CI: 0.0, 0.8) excess risk of hospitalization for IHD per
10 |_ig/m3 increase in PM25 two days earlier. The European APHEA-2 study (Le Tertre et al., 2002b)
considered PMi0 and found a 0.8% (95% CI: 0.3-1.2, lag 0-1) excess risk among those aged > 65 years.
Among the elderly in 5 cities in Australia and New Zealand (Barnett et al., 2006) there was a 4.3% (95%
CI: 1.9-6.4, lag 0-1) excess risk per 10 (ig/m3 increase in PM25. Among the elderly in several French
cities there was a 4.5% (95% CI: 2.3-6.8, lag 0-1) and 2.9% (95% CI: 1.5, 4.3, lag 0-1) excess risk per
10 (ig/m3 increase in PM2 5 and PM10, respectively (Host et al., 2007).
With regard to ED visits, the Atlanta-based SOPHIA study (Metzger et al., 2004) found positive
associations with PM10 and PM25 (ranging from 1.1 to 2.3%), but the effect estimates did not reach
statistical significance. In Sydney, Australia, Jalaludin et al. (2006) found a 0.8% (95% CI: -1.2 to 2.8%)
and 2.6% (95% CI: 0.1-5.2) excess risk of ED visits for IHD per 10 (ig/m3 increase in 2-day moving
average of PMi0 and PM2 5, respectively.
To explore this link further, Pope et al. (2006) used data from an ongoing registry of patients
undergoing coronary angiography at a single referral center in Salt Lake City, UT, between 1994 and
2004. The authors found a 4.8% (95% CI: 1.0-8.8, lag 0) excess risk of acute myocardial infarction or
unstable angina per 10 (ig/m3 increase in PM2 5 among 4818 patients. These results were robust to changes
in the definition of the outcome. The results of this study are particularly noteworthy given the high
specificity of the outcome definition.
In summary, large studies from the U.S., Europe, and Australia/New Zealand published since the
2004 PM AQCD provide support for an association between short-term increases in ambient levels of
PMio and PM2 5 and increased risk of hospitalization or ED visits for ischemic heart diseases. Estimates of
the excess risk vary considerably between studies, but as was the case for total CVD hospitalizations, the
excess risk appears to somewhat greater in Europe and Australia/New Zealand. Results from multicity
studies and U.S. and Canadian single city studies are presented in Figure 6-2.
December 2008
6-82
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Location Lag
Ischemic Heart Disease
Ito 2003, Lippmann et al. 2000
Detroit, Ml
1
Le Tertre et al. 2002
Europe 8 Cities
0-1
Metzger et al. 2004* (All Ages)
Atlanta, GA
0-2
Larrieu et al. 2007 (All Ages)
8 French Cities
0-1
Burnett et al. 1999 (All Ages)
Toronto, Canada
0-2
Le Tertre et al. 2002
8 European Cities
0-1
Jalaludin et al. 2006*
Sydney, Australia
0-1
Larrieu et al. 2007
8 French Cities
0-1
Ito 2003, Lippmann et al. 2000
Pope et al, 2006 (All Ages)
Host et al. 2007 (All Ages)
Metzger et al. 2004* (All Ages)
Barnett et al. 2006, (15-64 years)
Dominici et al. 2006
Barnett et al. 2006
Host et al. 2007
Burnett et al. 1999 (All Ages)
Ito 2003, Lippmann et al. 2000
Metzger et al. 2004* (All Ages)
Host et al. 2007 (All Ages)
Host et al. 2007 (65+ years)
Burnett et al. 1999 (All Aqes)
Detroit, Ml
1
Wasatch Front, UT
0
6 French Cities
0-1
Atlanta, GA
0-3
5 Cities Australia/NZ
0-1
204 US Counties
0
1
2
0-2 DL
5 Cities Australia/NZ
0-1
6 French Cities
0-1
Toronto, Canada
0-1
Detroit, Ml
Atlanta, GA
6 French Cities
6 French Cities
Toronto, Canada
1
0-3
0-1
0-1
0
PM,0
pmTTJ .
Myocardial Infarction
Linn et al. 2000 (>30 yrs)
Peters et al. 2001 (All Ages)
Peters et al. 2001 (All Ages)
Zanobetti & Schwartz (2005), 21
Peters et al. 2001 (All Ages)
Sullivan et al, 2005 (21-98 y)
Zanobetti & Schwartz 2005
Peters et al. 2001 (All Ages)
Los Angeles, CA
Boston, MA
Boston, MA
Boston, MA
King County WA
Boston, MA
Boston, MA
0
2 h
24 h
0
2 h
24 h
1 h
2 h
4 h
24 h
0
2 h
24 h
f%7]
PM„
II I I I I II I I I I II I I I I I II I I i I I II I I I
5 -1 2 5 8 11 15 19 23 27
Excess Risk Estimate
Figure 6-2. Excess risk estimates per 10 |jg/m3 increase in PM10, PM2 5, PWI10-25 for studies of ED
visits * and hospitalizations for IHD and Ml. Studies represented in the figure include
all multicity studies. Single city studies conducted in the U.S. and Canada are also
included.
December 2008
6-83
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
6.2.10.4. Acute Myocardial Infarction
Because even IHD refers to a heterogeneous collection of diseases and syndromes, several authors
have evaluated the association between short-term fluctuations in ambient PM and acute MI, a specific
form of ischemic heart disease.
In 2001, Peters et al. (2001) published their landmark study evaluating the effects of PM on the risk
of MI among 772 Boston-area participants in the Determinants of Myocardial Infarction Onset Study. The
authors found that a 10 |ig/m3 increase in the 2-h or 24-h average levels of PM2 5 was associated with a
17% (95% CI: 4-32) and 27% (95% CI: 6-53) excess risk of MI, respectively. In contrast, a similar study
among 5793 patients in King County, WA, found no association with PM2 5 with lag times of 1, 2, 4, or
24 h (Sullivan et al., 2005a). Among 852 hospitalized patients in Augsburg, Germany, Peters et al. (2005)
also found no association between PM2 5 and MI risk within this time frame, although they did find a
positive and statistically significant association with time spent in traffic (Peters et al., 2004).
These three studies are particularly important because in each one: (1) cases were prospectively
identified based on clinical criteria rather than retrospectively based on discharge diagnoses, and (2) time
of MI symptom onset was used for exposure assessment rather than date of hospital admission. Whether
the discrepant results among these studies are due to regional differences in population characteristics
and/or air pollution sources remains unclear. The King County study suggests that differences in statistical
approaches are unlikely to account for the discrepant results (Sullivan et al., 2005a). Analyses from the
U.S. MCAPS study suggest that substantial heterogeneity of effects are to be expected across regions of
the country (Bell et al., 2008a).
Several studies have assessed the association between acute exposure to ambient PM and MI using
administrative databases. In the U.S., MI was not one of the specific endpoints evaluated in the MCAPS
study (Dominici et al., 2006) or in the Atlanta-based SOPHIA study of ED visits (Metzger et al., 2004).
However, Zanobetti and Schwartz (2005) found a 0.7% (95% CI: 0.3-1.0) excess risk of MI per 10 (ig/m3
increase in same-day PMi0 among elderly Medicare beneficiaries in 21 cities. Subsequently, the same
authors found that among elderly Medicare beneficiaries living in the Boston metropolitan region a
10 |ig/m3 increase in PM25 was associated with a 4.9% (95% CI: 1.1, 8.2) excess risk on the same day
(Zanobetti and Schwartz, 2006).
This body of evidence may implicate traffic-related pollution generally as a risk factor for
myocardial infarction (MI). In the study described above, Peters et al. (2001) found positive associations
between risk of hospitalization for MI and potential markers of traffic-related pollution measured at a
central monitor including BC, CO and N02. However, none of these associations were statistically
significant in models adjusting for season, meteorological variables, and day of week. Zanobetti and
Schwartz (2006) examined the association between traffic-related pollution and risk of hospitalization for
December 2008
6-84
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
MI among Medicare beneficiaries in the Boston area and found that MI risk was positively and
significantly associated with measures of PM25, BC, N02, and CO, but not with levels of
non-traffic-related PM25. Peters et al. (2004) interviewed 691 subjects with MI who survived at least 24-h
after the event and found a strong positive association between self-reported exposure to traffic and the
onset of MI within one hour (OR: 2.9; 95% CI: 2.2, 3.8; P< 0.001). The association was somewhat
stronger among subjects traveling by bicycle or public transportation in the hour prior to the event. Of
note, however, this study did not directly measure traffic-related pollution.
Similar studies with administrative databases have been conducted in Europe, Australia, and New
Zealand. In Rome, D'Ippoliti et al. (2003) carried out a case-crossover study and found a statistically
significant positive association between total suspended particles (TSP) and the risk of hospitalization for
MI. Barnett et al. (2006) observed that in 5 cities in Australia and New Zealand, a 10 |ig/m3 increase in
PM2 5 was associated with a 7.3% (95% CI: 3.5, 11.4, lag 0-1 day) excess risk. In contrast, the HEAPSS
study found no evidence of an association between PMi0 and risk of hospitalization for a first MI in 5
European cities (2006a), although there is some indication that among survivors of a first myocardial
infarction, risk of re-hospitalization for MI may be related to transient elevations in PMi0 (von Klot et al.,
2005).
In summary, large studies from the U.S., Europe, and Australia/New Zealand published since the
2004 PM AQCD provide support for an association between short-term increases in ambient levels of
PM10 and PM2 5 and increased risk of hospitalization for MI, but the results have not been consistent.
These results need to be interpreted together with those studies evaluating hospitalization for IHD since
myocardial infarctions make up the majority of ischemic heart diseases. U.S. studies of MI are included in
Figure 6-2.
6.2.10.5. Congestive Heart Failure
Perhaps the first suggestion of an association between ambient PM and hospitalization for
congestive heart failure (CHF) was provided by the study of Schwartz and Morris (1995). These authors
reported that among elderly Medicare beneficiaries living in Detroit between 1986 and 1989, a 10 (.ig/nr1
increase inmeanPMio levels over the past two days was associated with a 1.0% (95% CI: 0.4, 1.6%)
increase in risk of hospitalization for CHF. As reviewed in the 2004 PM AQCD, using similar approaches,
statistically significant positive associations with PMi0 or PM2 5 were subsequently reported in single-city
studies looking at hospitalizations for CHF in Toronto (Burnett et al., 1999), Hong Kong (Wong et al.,
1999), and Detroit (Lippmann et al., 2000), but not Los Angeles (Linn et al., 2000) or Denver (Koken et
al., 2003).
December 2008
6-85
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Subsequent multicity studies support the presence of a positive association. In the U.S., Wellenius
et al. (2006) reported a 0.7% (95% CI: 0.4, 1.1) excess risk of hospitalization for CHF per 10 (ig/m3
increase in same-day PM10 among elderly Medicare beneficiaries in 7 cities. The larger MCAPS study
found a 1.3% (95%: 0.8, 1.8) excess risk per 10 |ig/m3 increase in same-day PM2.5 (Dominici et al., 2006).
In Australia and New Zealand, Barnett et al. (2006) found a 9.8% (95% CI: 4.8, 14.8, lag 0-1 day) and
4.6% (95% CI: 2.8, 6.3, lag 0-1 days) excess risk of hospitalization for CHF associated with a 10 (ig/m3
increase in PM2 5 and PMi0, respectively. Results from more recent single-city studies in Pittsburgh
(Wellenius et al., 2005b) and Taipei, Taiwan (Yang, 2008) have also reported positive associations.
Findings from the Atlanta-based SOPHIA study (Metzger et al., 2004) also support the presence of
a positive association. Specifically, the SOPHIA study found a 5.5% (95% CI: 0.6, 10.5, lag 0-2 days)
excess risk of emergency department visits for CHF per 10 |ig/m3 increase in the 3-day moving average of
PM2 5. Positive associations were also observed for CHF and EC and organic carbon components of
PM2 5. No associations were observed with PMi0.
Only one published study has attempted to evaluate the effects of ambient particles on CHF
symptom exacerbation using data which was not derived from administrative databases. Symons et al.
(2006) interviewed 135 patients with prevalent CHF hospitalized for symptom exacerbation in Baltimore,
MD. The authors found a 7.4% (95% CI: -7.5 to 24.2) excess risk of hospitalization per 10 |ig/m3 increase
in PM2 5 two days prior to symptom onset. Although the authors' findings did not reach statistical
significance, the study was ill-powered to find an effect of the expected magnitude.
In summary, large studies from the U.S., Europe, and Australia/New Zealand published since the
2004 PM AQCD provide support for an association between short-term increases in ambient levels of
PM10 and PM25 and increased risk of hospitalization and ED visits for heart failure. The excess risk
associate with heart failure hospitalizations and ED visits are consistently greater than those observed for
other CVD endpoints. The results of multicity studies and U.S. and Canadian single city studies are
summarized in Figure 6-3.
December 2008
6-86
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Location
Lag
Burnett et al. 1999
Toronto, Canada
0-2
Linn et al. 2000 (>30)
Los Angeles, CA
0
Ito 2003/Lippmann et al. 2000
Detroit, Ml
0
Metzger et al. 2004* (All Ages)
Atlanta, GA
0-2
Barnett et al. 2006 (15-64)
5 Cities Australia/NZ
0-1
Schwartz et al. 1995
Detroit, Ml
0-1
Morris et al. 1998
Chicago, IL
0
Wellenius et al. 2005
Pittsburgh, PA
0
Wellenius et al. 2006
7 US Cities
0
Barnett et al. 2006
5 Cities Australia/NZ
0-1
Burnett et al. 1999 (All Ages)
Toronto, Canada
0-2
Ito 2003/Lippmann et al. 2000
Detroit, Ml
1
Metzger et al. 2004* (All Ages)
Atlanta, GA
0-2
Barnett etal. 2006 (15-64 y)
5 Cities Australia/NZ
0-1
Symons et al. 2006 (All Ages)
Baltimore, MD
2
Dominici et al. 2006
204 US Counties
0
Barnett et al. 2006
5 Cities Australia/NZ
0-1
Burnett et al. 1999 (All Ages)
Ito 2003/Lippmann et al. 2000
Metzger et al. 2004* (All Ages)
Toronto, Canada
Detroit, Ml
Atlanta, GA
0-2
0
0-2
PMi
PM,
PM10-2.5
I I I I I I I I I I I I I I I
-6 -4 -2 0 2 4 6 8 10
Excess Risk Estimate
Figure 6-3. Excess risk estimates per 10 |jg/m3 increase in PM10, PM10-2.5 and PM10-2.5 for studies of
CHF ED visits * and hospitalizations. Studies represented in the figure include all
multicity studies. Single city studies conducted in the U.S. and Canada are also
included.
6.2.10.6. Cardiac Arrhythmias
1 A number of studies based on administrative databases have sought to evaluate the association
2 between short-term fluctuations in ambient PM levels and the risk of hospitalization for cardiac
3 arrhythmias (also known as dysrhythmias). Typically in these studies a primary discharge diagnosis of
4 ICD-9 427 has been used to identify hospitalized patients. However, ICD-9 427includes a heterogeneous
5 group of arrhythmias including paroxysmal ventricular or supraventricular tachycardia, atrial fibrillation
6 and flutter, ventricular fibrillation and flutter, cardiac arrest, premature beats, and sinoarterial node
7 dysfunction. One study in the Netherlands found that the positive predictive value of ICD-9 codes related
December 2008
6-87
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
to ventricular arrhythmias and sudden cardiac death was 82% (De Bruin et al., 2005). The positive
predictive value of other codes related to cardiac arrhythmias is unknown but likely to be lower.
The results from early studies of arrhythmia-related hospitalizations have been inconsistent with
positive findings in Toronto (Burnett et al., 1999) and null findings in Detroit (Schwartz and Morris,
1995), Los Angeles (Linn et al., 2000), and Denver (Koken et al., 2003). The U.S. MCAPS study found a
statistically significant 0.6% (95% CI: 0.0, 1.2%) excess risk of hospitalization for the combined end
point of cardiac arrhythmias and conduction disorders (ICD-9: 426, 427) per 10 (ig/m3 increase in
same-day PM2 5 (Dominici et al., 2006). A multicity study in Australia and New Zealand found no
evidence of an association between arrhythmia hospitalizations and either PMi0 or PM2 5 (Barnett et al.,
2006). Similarly, the Atlanta-based SOPHIA study found no evidence of an association between any
measure of ambient PM and the rate of ED visits for cardiac arrhythmias (Metzger et al., 2004).
A number of studies in patients with implanted cardioverter defibrillators have been more
successful at evaluating the link between ambient air pollution and the risk of atrial and ventricular
arrhythmias (Berger et al., 2006; Dockery et al., 2005a; Dockery et al., 2005b; Peters et al., 2000; Rich et
al., 2004; Vedal et al., 2004). An important strength of these studies is the ability to examine recordings of
arrhythmic episodes, thereby reducing misclassification of the outcome. These studies are reviewed in
detail in Section 6.2.2.1.
In summary, the current evidence does not support the presence of a consistent association between
short-term increases in ambient levels of PM10 and PM2 5 and increased risk of hospitalization for cardiac
arrhythmias. However, it should be noted that studies of hospital admissions or ED visits are ill-suited to
the study of cardiac arrhythmias. Studies in patients with implanted defibrillators, human panel studies
with ambulatory ECG recordings, and animal toxicological studies provide a more appropriate setting for
evaluating this endpoint.
6.2.10.7. Cerebrovascular Disease
Time-series studies evaluating the hypothesis that short-term increases in ambient PM levels are
associated with increased risk of hospitalization for cerebrovascular disease have been inconsistent with a
minority of studies reporting statistically significant positive associations (Chan et al., 2006a; Dominici et
al., 2006; Metzger et al., 2004; Wordley et al., 1997), and several studies reporting null or negative
associations (Anderson et al., 2001; Barnett et al., 2006; Burnett et al., 1999; Jalaludin et al., 2006;
Larrieu et al., 2007; Le Tertre et al., 2002b; Peel et al., 2007; Villeneuve et al., 2006; Wong et al., 1999).
The U.S. MCAPS study found a 0.8% (95% CI: 0.3, 1.4) excess risk of hospitalization for
cerebrovascular disease per 10 (ig/m3 increase in same-day PM2 5 (Dominici et al., 2006). Interestingly,
the association showed regional variability with the strongest associations observed in the Eastern U.S.
December 2008
6-88
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
The Atlanta-based SOPHIA study found a 2.0% (95% CI: -0.1, 4.3, lag 0-2 days) excess risk of ED visits
for a combined endpoint of cerebrovascular and peripheral vascular disease excluding hemorrhagic
strokes per 10 (ig/m3 increase in PM10 and a 5.0% (95% CI: 0.8, 9.3, lag 0-2 days) excess risk for PM25
(Metzger et al., 2004).
In contrast, large multicity studies outside of North America have failed to observe an association.
The APHEA study, found a 0.0 % (95% CI: -0.3, 0.3) excess risk of hospitalization for cerebrovascular
disease per 10 |ig/m3 increase in the 2-day moving average of PMi0 in 8 European cities (Le Tertre et al.,
2002b). Investigators from the French PSAS program reported a 0.8% (95% CI: -0.9, 2.5, lag 0-1 days)
excess risk per 10 (ig/m3 increase in PMi0 among patients aged > 65 years and a 0.2% (95% CI: -1.6, 1.9,
lag 0-1 days) excess risk among all patients (Larrieu et al., 2007). Although neither estimate was
statistically significant, the estimated excess risk among the elderly is very similar to that observed in the
U.S. MCAPS study. Barnett et al. (2006) examined this hypothesis in New Zealand and Australia and
found no association, but the authors did not report point estimates or confidence intervals.
All of the above studies have identified cases of cerebrovascular disease based on ICD-9 or ICD-10
codes (most commonly ICD-9 430-438). However, the range of ICD codes commonly used in these
studies includes ischemic strokes, hemorrhagic strokes, transient ischemic attacks (TIAs) and several
poorly defined forms of acute neurological events (e.g. seizures from a vascular cause) (see Table 6-5). It
is plausible that ambient PM has different effects on each of these disparate outcomes.
6.2.10.8. Ischemic Strokes and Transient Ischemic Attacks
An increasing number of studies have specifically evaluated the association between PM and the
risk of ischemic stroke (Chan et al., 2006a; Henrotin et al., 2007; Linn et al., 2000; Lisabeth et al., 2008;
Tsai et al., 2003b; Villeneuve et al., 2006; Wellenius et al., 2005a). Linn et al. (2000) found a 1.3% (95%
CI: 1.0, 1.6 per 10 |ig/nr\ lag 0) excess risk of hospitalization for ischemic stroke in the Los Angeles
metropolitan area. Wellenius et al. (2005a) reported a statistically significant 0.4% (95% CI: 0.0, 0.9)
excess risk per 10 (ig/m3 increase in same-day PMi0 among elderly Medicare beneficiaries in 9 U.S.
cities. In Kaohsiung, Taiwan, Tsai et al. (2003b) found a 5.9% (95% CI: 4.3, 7.4, lag 0-2 days) excess risk
of hospitalization for ischemic stroke per 10 (ig/m3 increase in PMi0 after excluding days with mean daily
temperature < 20°C. Meanwhile, in Taipei, Taiwan, Chan et al. (2006a) found a 1.6% (95% CI: -0.8, 3.9,
lag 3) and 3.0% (95% CI: -0.8, 6.6, lag 3) excess risk per 10 (ig/m3 increase in PMi0 and PM25,
respectively. Villeneuve et al. (2006) found no association between either PM2 5 or PMi0 and emergency
department visits for acute ischemic stroke in Edmonton, Canada.
Two recent studies are particularly noteworthy given the high specificity of the outcome definition.
Henrotin et al. (2007) used data on 1432 confirmed cases of ischemic stroke from the French Dijon Stroke
December 2008
6-89
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Register and found 0.9% (-7.0, 9.4%) excess risk of ischemic stroke per 10 |ig/m3 increase in PM10 on the
same day and a 1.1% (-0.2, 9.4%) excess risk on the previous day (lag 1 day). Lisabeth et al. (2008) used
data on 2350 confirmed cases of ischemic stroke and 1158 cases of TIA from the Brain Attack
Surveillance in Corpus Christi Project (BASIC), a population-based stroke surveillance project designed
to capture all strokes in Nueces County, Texas. The authors found a 6.0% (95% CI: -0.8, 13.2) and 6.0%
(95% CI: -1.8, 14.4) excess risk of ischemic stroke/TIA per 10 (ig/m3 increase in PM2 5 on the previous
day and the same day, respectively.
Only the study by Villeneuve et al. (2006) specifically evaluated the association between ambient
PM and the risk of TIAs. This study failed to find any evidence of an association with either PM2 5 or
PM10.
Hemorrhagic Strokes
Most of the studies in the preceding section also evaluated the association between ambient PM
and the risk of hemorrhagic stroke (Chan et al., 2006; Henrotin et al., 2007; Tsai et al., 2003; Villeneuve
et al., 2006; Wellenius et al., 2005). In Kaohsiung, Taiwan, Tsai et al. (2003) noted a 6.7% (95% CI: 4.2,
9.4, lag 0-2 days) excess risk of hospitalization for hemorrhagic stroke per 10 (.ig/nr1 increase in PMi0,
after excluding days where the mean temperature was < 20°C. However, in the U.S., Wellenius et al.
(2005) failed to find any association between ambient PMi0 levels and risk of hemorrhagic stroke among
Medicare beneficiaries in 9 U.S. cities. Similarly, Villeneuve et al. (2006) found no evidence of an
association between emergency department visits for hemorrhagic stroke and either PM10 or PM2 5 levels
in Edmonton, Canada. Henrotin et al. (2007) found no evidence of an association between risk of
hospitalization and PM10 levels in Dijon, France, and Chan et al. (2006) found no evidence of an
association between risk of hospitalization and either PM10 or PM2 5 levels in Taipei, Taiwan.
In summary, large studies from the U.S., Europe, and Australia/New Zealand published since the
2004 PM AQCD provide inconsistent support for an association between short-term increases in ambient
levels of PM10 and PM2 5 and risk of hospitalization and ED visits for cerebrovascular disease (Figure
6-4). The heterogeneity in results is likely partly attributed to differences in the sensitivity and specificity
of the various outcome definitions used in the relevant studies. Effect estimates from multicity studies and
single city U.S. and Canadian studies are included in Figure 6-4.
December 2008
6-90
DRAFT-DO NOT QUOTE OR CITE
-------
Legend
CBVD Cerebrovascular Disease
PVD Peripheral Vascular Disease
IS Ischemic Stroke
TIA Transient Ischemic Attack
HS Hemorrhagic Stroke
Reference
Location
Lag
Metzger et al (2004)* (All Ages)
Atlanta, GA
0-2
Le Tertre (2002)/reanalysis
8 European Cities
0-1
Larrieu et al. 2007
8 French Cities
0-1
Larrieu et al. 2007 (All Ages)
8 French Cities
0-1
Dominici et al. 2006
204 US Counties
0
Northeastern US
Southeastern US
Midwestern US
Southern US
Metzger et al. 2004 (All Ages)
Atlanta, GA
0-2
Wellenius et al. 2005b
Lisabeth et al. 2008 (All Ages)
Villeneuve et al. 2006,
Linn etal. 2000(>30yrs)
Villeneuve et al. 2006
9 US Cities
Nueces County TX
Edmonton, Canada
Los Angeles, CA
Edmonton, Canada
Wellenius 2005b
Villeneuve et al. 2006
Villeneuve et al. 2006
9 US Cities
Edmonton, Canada
Edmonton, Canada
0
1
0
0-2
0-2
0-2
0-2
0
0-2
0-2
0-2
0-2
0
0-2
0-2
0-2
0-2
PM
10
PM
2.5
PM
10
PM2.5
PM
10
PM
2.5
—CBVD & PVD
f CBVD
-CBVD
CBVD
CBVD
-CBVD
CBVD
CBVD
— CBVD
- CBVD & PVD
. IS/TIA
— IS/TIA
-IS .Cool
IS, Warm
TIA, Cool
TIA, Warm
IS , Cool
IS, Warm
TIA , Cool
. TIA , Warm
-HS
HS , Cool
HS, Warm
HS, Warm
i i i i i i i i i i i i i i i i i i i i
-20 -14 -8 -2 2 6 10 14 18 22
Excess Risk Estimate
Figure 6-4. Excess risks estimates per 10 |jg/m3 increase in PM10, PM2.5, and PM10-2.5 for studies of
ED visits* and hospitalizations for cerebrovascular diseases. Studies represented in
the figure included all multicity studies. Single city studies conducted in the U.S. and
Canada are also included.
December 2008
6-91
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Peripheral Vascular Disease
In the U.S., the large MCAPS study (Dominici et al., 2006) evaluated the association between
mean daily PM2 5 levels and the risk of hospitalization among elderly Medicare beneficiaries in 204 urban
counties and found that a 10 (ig/m3 increase in PM2 5 was not significantly associated with risk of
hospitalization for peripheral vascular disease 0-2 days later. An earlier study in Toronto (Burnett et al.,
1999) found a negative association with PM2 5 (point estimate and confidence intervals not reported), a
positive non-significant association with PMi0 (0.5%; 95% CI: -0.5, 1.6%), and a positive statistically
significant association with PM10-2.5 (2.2%; 95% CI: 0.1, 4.3%). The Atlanta-based SOPHIA study
(Metzger et al., 2004) of emergency department visits grouped visits for PVD with those for
cerebrovascular disease, making interpretation of these results challenging.
In summary, there is insufficient published data to determine whether or not there may be an
association between short-term increases in ambient levels of PMi0 and PM2 5 and increased risk of
hospitalization and ED visits for peripheral vascular disease.
PM and Out of Hospital Cardiovascular Deaths
One study of out of hospital cardiac death conducted in Seattle, WA (Checkoway et al., 2000),
which reported no association with PM was included in the 2004 PM AQCD. In the U.S., the survival rate
of sudden cardiac arrest is less than 5%. In addition, as discussed above, Zeka et al. (2006a) found that the
estimated mortality risk due to short-term exposure to PMi0 was much higher for out-of-hospital
cardiovascular deaths than for in-hospital cardiovascular deaths. The analysis of studies that examine the
association between PM and cardiac arrest could provide evidence for an important link between the
morbidity and mortality effects attributed to PM.
Sullivan et al. (2003) examined the association between the incidence of primary cardiac arrest and
daily measures of PM2 5 (measured by nephelometry) using a case-crossover analysis of 1,206
Washington State out-of-hospital cardiac arrests (1985-1994) among persons with (n = 774) and without
(n = 432) clinically recognized heart disease. The authors examined PM associations at 0- through 2-day
lags using the time-stratified referent sampling scheme (i.e., the same day of the week and month of the
same year). The estimated relative risk for a 13.8-(.ig/ni3 increase in 1-day lag PM25 (nephelometry: 0.54 x
10-1 km-1 bsp) was 0.94 (95% CI: 0.88-1.02), or 0.96 (0.91, 1.0) per 10 (ig/m3 increase. Similar
estimates were reported for 0- or 2-day lags. The presence or absence of clinically recognized heart
disease did not alter the result. This finding is consistent with the previous study of cardiac arrest in
Seattle (Levy et al., 2001) that reported no PM association. It is also consistent with the Sullivan et al.
(2005a) analysis of PM and onset of MI, and the Sullivan et al. (2007) analysis of PM and blood markers
of inflammation in the elderly population, both of which were conducted in Seattle. Note also that the
December 2008
6-92
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
analysis of the NMMAPS data for the years 1987-1994 also found no PM10 association for all-cause
mortality in Seattle. Overall, the results of studies conducted in Seattle consistently found no association
between PM and cardiovascular outcomes or all-cause mortality.
Rosenthal et al. (2008) examined associations between PM2.5 and out-of-hospital cardiac arrests in
Indianapolis, Indiana for the years 2002-2006 using a case-crossover design with time-stratified referent
sampling. Using all the cases (n = 1,374), they found no associations between PM2 5 and cardiac arrest in
any of the 0 through 3-day lags or multi-day averages thereof (e.g., for 0-day lag, OR = 1.02 [CI: 0.94,
1.11] per 10 |_ig/m3 increase in PM25). However, for cardiac arrests witnessed by bystanders (n = 511),
they found a significant association with PM2 5 exposure (by TEOM, corrected with FRM measurements)
during the hour of the arrest (OR = 1.12 [CI: 1.01, 1.25] per 10 (ig/m3 increase in PM25). and even larger
risk estimates for older adults (age 60-75) or those that presented with asystole. There have been very few
PM studies that used hourly PM measurements, and further studies are needed to confirm associations at
such time scales.
In Rome, Forastiere et al. (2005) examined associations between air pollution (particle number
concentration, or PNC, PMi0, CO, N02, and 03) and out-of-hospital coronary deaths (n = 5,144) for the
study period of 1998-2000. A case-crossover design with the time-stratified referent sampling was used to
examine the pollution indices at lag 0 through 3 days and the average of 0-1 lags. They found associations
between deaths and PNC (lag 0 and 0-1), PM10 (lag 0, 1, and 0-1), and CO (lag 0 and 0-1) but not with
N02 or 03. The risk estimate for 0-day lag PM10 was 1.59% (CI: 0.03, 3.18) per 10 (ig/m3 increase. The
older adults (65-74 and 75+ age groups) showed higher risk estimates than the younger (35-64) age
group. Because PNC is considered to be associated with ultrafine particles, and CO was also associated
with out-of-hospital cardiac arrests, combustion sources were implicated.
In summary, only a few studies have examined out-of-hospital cardiac arrest or deaths. The two
studies from Seattle, WA consistently found no association (also consistent with other cardiac effects and
mortality studies conducted in that locale); a study in Indianapolis, IN found an association with hourly
PM2 5 but not daily PM2 5; and a study in Rome found an association with PMi0 but also with particle
numbers and CO. Because multicity mortality studies examining this association found heterogeneity in
PM risk estimates across regions, future studies of out-of-hospital cardiac arrest will need to consider
location and the air pollution mixture during their design.
December 2008
6-93
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
6.2.11. Summary of Causal Determinations by PM Metric
6.2.11.1. PM10
Causal Determination
The bulk of recent evidence for PMi0 health effects is derived from epidemiologic studies of
hospital admissions and ED visits. Although some regional heterogeneity is evident in the single city
effect estimates, consistent increases in hospital admission or ED visits for cardiovascular diseases have
been observed. Further, multicity analyses indicate a significant increase in CVD admissions that are
consistent with those reported in the 2004 PM AQCD. The recent literature also implicates IHD and CHF
as largely responsible for these CVD admissions rather than cerebrovascular diseases. However, a large
multicity U.S. study provides evidence of an association of PMi0 with ischemic stroke. The results of
these studies provide support of associations between short-term PMi0 exposure and increased risk of
cardiovascular admissions in areas with mean concentrations from 16.8 to 48 (ig/m3 (Figure 6-1).
No human clinical studies have evaluated the effect of PMi0 on measures of cardiovascular
function. However, animal toxicological studies have been published since the 2004 PM AQCD that
employ PMi0, demonstrating impacts on the cardiovascular system. A recent inhalation study found
lowered myocardial contractility upon exposure to PMi0, while several intratracheal instillation studies
observed altered vasoreactivity and elevated levels of systemic inflammatory and blood coagulation
markers, thus providing coherence with the epidemiologic findings of increases in hospital admissions
and ED visits. Several epidemiologic studies of HRV, systemic markers of inflammation and coagulation
as well as oxidative stress demonstrated PM10-related effects in areas with mean concentrations ranging
from 14 to 42 |ig/nr\ Coherence of specific endpoints across epidemiologic and toxicological studies is
strongest for vasomotor function and coagulation markers.
Overall, recent evidence supports the conclusion of the 2004 PM AQCD that short-term exposure
to PMio is associated with an increased risk of cardiovascular morbidity. Furthermore, findings of altered
systemic inflammation, autonomic function, coagulation, and vasoreactivity provide biological
plausibility that exposure to PMi0 could lead to more severe effects, including hospital admissions or ED
visits for ihd, CHF, or ischemic stroke. Collectively, the evidence is sufficient to conclude that a
causal relationship is likely to exist between PM10 exposures and cardiovascular morbidity
Heart Rate Variability
Epidemiologic Studies: Discrepant findings across six epidemiologic studies of HRV were
described in the 2004 PM AQCD. Although approximately 20 new studies of HRV have been conducted
December 2008
6-94
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
only three included ambient PM10 concentration (Ebelt et al., 2005; Liao et al., 2005; Lipsett et al., 2006).
Findings from these studies were consistent, with each reporting decreases in time and frequency domain
measures of HRV.
Arrhythmia
Epidemiologic Studies: Recent studies of PMio and ventricular arrhythmias report no associations
(Metzger et al., 2004; Rich et al., 2004; Vedal et al., 2004). However, Ebelt et al. (2005) reported an
increase in supraventricular ectopic beats with same day PMi0 concentration.
Vasomotor Function
Epidemiologic Studies: One recent study investigating the association of PMi0 with vasomotor
function reported increased FMD, but decreased basal diameter in healthy volunteers (Liu et al., 2007b).
Toxicological Studies: There were no studies that directly measured vasomotor function in the
2004 PM AQCD. Endothelin, an activator of vasoconstriction, was evaluated in two studies that reported
increased plasma ET-1 and ET-3 levels in rats following extremely high exposure concentrations to PMi0
(EHC-93). There are two intratracheal instillation studies that employed PMi0 (EHC-93) to evaluate
vasoreactivity, although over 99% of the particles were < 3 |im. One study in carotid arteries of rabbits
exposed to PMi0 reported reduced ACh-stimulated relaxation, whereas the other study in aortic rings of
SH rats found enhanced ACh-induced relaxation. No change in vasoconstriction was observed in the latter
study and it was not measured in the former study.
Blood Pressure
Epidemiologic Studies: Findings from the previous AQCD were inconsistent regarding the
association of PMi0 (or TSP) with BP. No increases in BP were reported in recent studies examining PMi0
(Ebelt et al., 2005; Jansen et al., 2005).
Toxicological Studies: Only one animal toxicological study reviewed in the 2004 PM AQCD
measured BP and did not report any effect following exposure to diesel soot (PM10). There are no new
studies that evaluated BP responses and exposure to PM10 at reasonable concentrations.
Cardiac Contractility
Toxicological Studies: A recent inhalation study conducted in mice used echocardiography and
demonstrated reductions in cardiac fractional shortening, diminished ejection fraction, and maximum
change in pressure overtime with carbon black exposure. These results support lowered myocardial
contractility.
December 2008
6-95
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Systemic Inflammation
Epidemiologic Studies: Several studies of PMio (and TSP) reviewed in the 2004 PM AQCD
provided limited support for mechanisms related to CVD development or progression. Newer studies
have focused on PM2 5. Findings in the few new studies examining PM10 have been inconsistent (Ruckerl
et al., 2006; 2007b; 2008; Sullivan et al., 2007).
Toxicological Studies: There were mixed results reported in the 2004 PM AQCD in regard to
systemic inflammation. Only one toxicological study conducted in rabbits demonstrated increased
circulating PMN band cell counts following PMi0 exposure (EHC-93). The one recent study that
employed intratracheal instillation of PMi0 reported elevated systemic IL-6 and TNF-a levels.
Blood Coagulation
Epidemiologic Studies: Studies reviewed in the 2004 PM AQCD reported associations of PMi0
with blood viscosity and fibrinogen. Newer studies have focused on PM2 5; however, PMi0 was associated
with an increase in vWF, but not fibrinogen in a U.S. multicity study (Liao et al., 2005), fibrinogen in a
European multicity study (Ruckerl et al., 2007b) and prothrombin time (Baccarelli et al., 2007a).
Toxicological Studies: A recent study reported numerous blood coagulation measures that were
altered in mice exposed to PMi0 from Dusseldorf, Germany, which indicated accelerated coagulation.
Cardiac or Systemic Oxidative Stress
Epidemiologic Studies: Two new studies demonstrated associations between PMi0 and oxidative
stress measurements of TBARS and tHcy (Baccarelli et al., 2007a; Liu et al., 2007b).
Clinical Outcomes in Epidemiologic Studies
An increase in CVD admissions with PMi0 concentration between 0.6% and 1.7%, per 10 |ig/m3
was estimated based on studies reviewed in the 2004 PM AQCD. Pooled estimates from MCAPS and
other multicity analyses indicate a significant increase in admissions within this range (Barnett et al.,
2006; Larrieu et al., 2007; Le Tertre et al., 2002b; Zanobetti and Schwartz, 2003). Heterogeneity in effect
estimates is evident across single city studies. Newer studies have indicated that IHD and CHF rather than
cerebrovascular diseases appear to drive these associations observed between PM10 and cardiovascular
disease admissions and ED visits. Out-of-hospital cardiac arrest was associated with PM10 in one study
conducted in Rome.
December 2008
6-96
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
6.2.11.2. PMio-2.5
Causal Determination
Several recent epidemiologic studies of the effect of ambient PM10-2.5 concentration on hospital
admissions and ED visits for cardiovascular diseases have been conducted. Of particular note, a study of
Medicare patients in 108 counties across the U.S. reports that PM10_2.5 is not associated with
cardiovascular disease admissions after adjustment for PM2 5. There is a small body of evidence from
single city studies and a 6-city study in France that may provide evidence to the contrary, but the U.S.
study of Medicare patients is the only study to adjust PM10-2.5 for PM25.
There are very few studies that examined the effect of exposure to PMi0_2.5 on cardiovascular
endpoints or biomarkers in humans or animals. Two human clinical studies found decreases in HRV in
healthy subjects following exposure, and a trend toward an increase in blood coagulation factors was also
demonstrated in one of these studies. The only PM10-2.5 toxicological studies that evaluated cardiovascular
responses were comparative studies of various size fractions and only blood or plasma parameters were
measured. Both studies used non-inhalation methodologies and relatively high doses of PM. Therefore,
evidence of biological plausibility for cardiovascular morbidity effects following PMi0.2.5 exposure is
sparse.
Limited evidence exists for short-term PMi 0-2.5 exposures and cardiovascular morbidity. In its
entirety, the literature shows that evidence is inadequate to determine if a causal relationship exists
between PM10-2.5 exposures and cardiovascular morbidity.
Heart Rate Variability
Epidemiologic Studies: Findings from five recent studies of the effect of PM10_2 5 on HRV were
inconsistent (Adar et al., 2007a; Ebelt et al., 2005; Lipsett et al., 2006; Timonen et al., 2006; Yeatts et al.,
2007). Study populations yielding these discrepant results included bus riders, asthmatics, COPD and
heart disease patients.
Human Clinical Studies: Two new studies have observed decreases in HRV (SDNN) among
healthy adults, but not asthmatics, following exposure to PM10_2.5 (Gong et al., 2004b; Samet et al., 2007).
Arrhythmia
Epidemiologic Studies: The only recent study that examined the effect of PM10-2.5 on ICD
recorded arrhythmias (Metzger et al., 2007) reported null findings. However, an association of PM10.2.5
was observed with supraventricular ectopic beats (Sarnat et al., 2006c).
December 2008
6-97
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Ischemia and ECG Abnormalities Indicating Ischemia
Epidemiologic Studies: One recent study examined PMi0.2.5 in relation to ST-segment depression
and found an association that was comparable to those observed with other PM metrics (e.g. PM2 5, UFP
and ACP) (Pekkanen et al., 2002).
Blood Pressure
Epidemiologic Studies: In the only study of PMio.2.5 and BP, decreases were reported (Ebelt et al.,
2005).
Blood Coagulation
Human Clinical Studies: One new study reported a trend toward elevated levels of pro-clotting
factors following exposure to thoracic coarse CAPs in a group of healthy adults (Samet et al., 2007).
Clinical Outcomes in Epidemiologic Studies
Two studies of the association of PMi 0-2.5 with cardiovascular admissions were reviewed in the
2004 PM AQCD. The recent study by Peng et al. (2008) provides new insights regarding the relative
toxicity of PM10-2 5 versus PM2 5. After adjustment for PM2 5, there was no association between PM10-2.5
and risk of CVD hospitalizations. In the SOPHIA study associations of cardiovascular outcomes with
PM10-2 5 were weak and not statistically significant compared to associations with levels of PM2 5.
Similarly the French PSAS program reported substantially weaker associations for PM10-2.5 and total CVD
and cardiac hospitalizations. However, these investigators reported similar effect sizes per IQR increases
in PM2.5 and PM10-2.5 for the outcome of IHD. Results from single city studies are heterogenous and
imprecise.
6.2.11.3. PM2.5
Causal Determination
A large body of evidence from studies of the effect of PM2 5 on hospital admissions and ED visits
for cardiovascular diseases has been published since the 2004 PM AQCD. Associations with PM2 5 are
consistently positive with the majority of studies reporting increases in hospital admissions or ED visits
ranging from a 0.5 to 3.4% per 10 |ig/m3 increase in PM2 5. The largest U.S.-based multicity study,
Medicare Air Pollution Study (MCAPS) reported excess risks in the range of approximately 0.7% with
the largest excess risks in the North East (1.08%) and in the winter (1.49%). These PM2 5 effects appear to
be driven by IHD and CHF rather than cerebrovascular diseases. Several recent studies have examined
associations of specific PM2 5 components or sources. Positive associations have been observed with
December 2008
6-98
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
several components produced by traffic or combustion and source apportionments have linked these
sources to cardiovascular hospital admissions and ED visits. However, no single component has been
identified to explain the toxicity of PM25. Overall, the results of these studies provide support of
associations between short-term PM2 5 exposure and increased risk of cardiovascular admissions in areas
with mean concentrations ranging from 13.8 to 18.8 |ig/nr\ Numerous epidemiologic studies of HRV,
ECG abnormalities, vasomotor function, systemic inflammation, coagulation and oxidative stress support
the biological plausibility of these effects at ambient levels.
Changes in various measures of cardiovascular function have been consistenly demonstrated
following controlled human exposures to PM2 5. The majority of the new studies described have been
conducted using DE or CAPs, and provide strong evidence of PM2 -induced decreases in HRV and
vasomotor function, as well as increases in markers of systemic oxidative stress. One new study also
observed a decrease in ST-segment depression following exposure to DE in a group of older adults with
prior MI. Although not consistently observed across studies, some investigators have reported
PM2 5-induced changes in BP, blood coagulation markers, and markers of systemic inflammation. These
findings from human clinical studies provide coherence and biological plausibility for the associations
observed in epidemiologic studies.
A number of toxicological studies have been conducted that demonstrate findings with PM2 5 and
cardiovascular endpoints. Consistent with evidence from human clinical studies, the most significant
contributions from the current toxicological literature for acute PM2 5-induced cardiovascular effects are
decreased myocardial blood flow following ischemia, changes in vascular reactivity and morphology, and
increased cardiac oxidative stress. Results for HRV, arrhythmia, systemic inflammation, and blood
coagulation are mixed. For BP and cardiac contractility, very few studies were evaluated or the study
design was weak.
Together, the collective evidence is sufficient to conclude that there is a causal relationship
between relevant PM2.5 exposures and cardiovascular morbidity
Heart Rate Variability
Epidemiologic Studies: Discrepant findings across six epidemiologic studies of HRV were
described in the 2004 PM AQCD. Overall, the majority of recent studies have observed decreases in time
and frequency domain measures of HRV with PM2 5 (findings for LFHFR were not consistent across
studies). However, heterogeneity in effects was observed within the U.S. and between 3 European cities
studies (Timonen et al., 2006). Several studies included PM25 components including BC, EC, sulfate and
secondary PM in their analyses (Adar and Kaufman, 2007; Ebelt et al., 2005; Luttmann-Gibson et al.,
2006; Park et al., 2005b; Schwartz et al., 2005b; Wheeler et al., 2006a). Although some studies showed
December 2008
6-99
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
decreases in time or frequency domain HRV measures associated with these components, results were not
consistent. Finally, several analyses from the Normative Aging Study in Boston as well as a randomized
trial of omega-3 fatty acid in Mexico City have indicated a role for oxidative stress in the relationship
between the PM2 5 and HRV (Chahine et al., 2007; Romieu et al., 2005)Park et al. 2005; Park et al. 2008;
Park et al. 2006; Schwartz et al. 2005b).
Human Clinical Studies: Two pilot studies cited in the 2004 PM AQCD found no effect of PM2 5
CAPs on HRV in healthy adults (Gong et al., 2000; Petrovic et al., 2000). However, a larger study
reported a significant decrease in the ratio of LF/HF power in healthy and asthmatic adults immediately
following a 2-h exposure to fine CAPs (Gong et al., 2003a). Two more recent human clinical studies have
demonstrated a decrease in HRV (SDNN and PNN50) among healthy older adults, but not older adults
with COPD, following exposure to PM25 CAPs (Devlin et al., 2003; Gong et al., 2004a). Another new
human clinical study reported some evidence of a decrease in LF/HF power ratio 1-h following exposure
to DE in healthy subjects and subjects with metabolic syndrome (Peretz et al., 2008b).
Toxicological Studies: There were two animal toxicological studies in the 2004 PM AQCD that
examined HRV. Dogs exposed to CAPs had decreased HR and increased HRV and rats with induced MI
had lowered SDNN compared to those exposed to air or carbon black. In the latest PM2 5 CAPs studies,
increase and decreases in H were observed. For the carbon black studies that examined H, bradycardia
was reported. Fine sulfuric acid did not induce any H effect, but DE induced lowered H in ApoE" " mice.
Increased HRV measures (SDNN, LF/HF ratio, and rMSSD) were observed with carbon black exposure.
Diesel exposure resulted in decreased rMSSD in rats with congestive heart failure. Using source
apportionment methodologies, H and HRV changes were associated with resuspended soil, secondary
sulfate, residual oil, and motor vehicle/other sources. In a separate research article from this study, Ni was
implicated in elevated H and lowered SDNN.
Arrhythmia
Epidemiologic Studies: The initial study indicating a possible association of PM25 and BC with
ventricular arrhythmias among a cohort of patients with ICDs in Boston was reviewed in the 2004 PM
AQCD (Peters et al., 2000). Several additional analyses of this same cohort that examined different lags
and pollutants have also reported increased risk of arrhythmia with PM2 5, BC and sulfate exposure
(Dockery et al., 2005a; 2005b). Studies conducted in other regions (e.g. St. Louis, Vancouver, Atlanta)
have not supported these findings (Dusek et al., 2006; Metzger et al., 2004; Rich et al., 2004; Vedal et al.,
2004). However, three studies of supraventricular ectopic beats or supraventricular tachycardia conducted
in regions outside of the North East have all shown positive associations with PM2 5 (Berger et al., 2006;
Ebelt et al., 2005; Sarnat et al., 2006) and sulfate in one study (Sarnat et al., 2006). PM25, ACP (0.1-1 |im)
December 2008
6-100
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
EC and OC, traffic and combustion particles were associated with repolarization parameters in two
analyses of IHD patients in Erfurt Germany (Henneberger, 2005; Yue et al., 2007).
Toxicological Studies: Arrhythmogenesis was reported for studies reviewed in the 2004 PM
AQCD. Generally these results were observed in animal models of disease (SH rat, MI, pulmonary
hypertension) exposed to non-atmospheric PM2 5 (i.e., ROFA, DE, metals). One recent study employing
PM2 5 CAPs demonstrated decreased arrhythmia frequency in a rodent model of MI. In contrast, older rats
had elevated frequency of delayed beats with PM2 5 CAPs exposure. Increased incidence of VPB was
reported with DE in a rat model of CHF and in ApoE_/" mice, although for the latter study, it appeared that
the gases were responsible for the effect. For a study employing gasoline exhaust, the particle fraction
was required to induce T-wave changes in ApoE_/" mice.
Ischemia
Epidemiologic Studies: Two recent studies of ST-segment depression and PM2 5 report adverse
associations (Gold et al., 2005; Pekkanen et al., 2002). Gold et al. (2005) also reported an association with
BC and Pekkanan et al. (2002) reported an association with ACP (0.1-1 |im).
Human Clinical Studies: One recent study observed an increase in exercise-induced ST-segment
depression during exposure to DE in a group of subjects with prior MI (Mills et al., 2007).
Toxicological Studies: In the 2004 PM AQCD, one study reported increased ST-segment
magnitude in response to PM2 5 CAPs in dogs with experimentally-induced ischemia. However, in dogs
with signs of naturally occurring heart disease, ROFA did not induce any change in ST-segment or the
T-wave. A recent study of dogs with induced ischemia reported increased ST-segment elevation with
PM2 5 CAPs exposure that was linked to the mass concentration of Si as a tracer of source. Myocardial
blood flow during myocardial ischemia was decreased and coronary vascular resistance was increased in
dogs following PM2 5 CAPs exposure.
Vasomotor Function
Epidemiologic Studies: Studies of fine PM (e.g. PM2 5, PMi) (Dales et al., 2007; O'Neill et al.,
2005b; Rundell, 2007) and components (e.g. PM2 5, PMi, sulfate, BC) (O'Neill et al., 2005b) have been
conducted since the 2004 PM AQCD. Decreases in FMD and BAD were observed with fine PM and BC
in healthy and diabetic populations.
Human Clinical Studies: The 2004 PM AQCD presented the results of one study that observed a
decrease in BAD following exposure to fine CAPs in combination with ozone (Brook et al., 2002). A
subsequent study found that this effect on vasoreactivity was associated with both the organic and EC
fraction of the fine CAPs (Urch et al., 2004). One additional recent study found no effect of fine CAPs,
December 2008
6-101
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
composed largely of sea salt, on vasomotor function in healthy adults (Mills et al., 2008). However,
several new human studies have observed decreases in forearm blood flow and BAD following exposure
to DE or fine indoor air particles (Brauner et al., 2008; Mills et al., 2005; Peretz et al., 2008b; Tornqvist et
al., 2007).
Toxicological Studies: There were no studies that directly measured vasomotor function in the
2004 PM AQCD. Endothelin, an activator of vasoconstriction, was evaluated in one study that reported
increased plasma ET-3 levels in rats following exposure to diesel PM. Six new studies evaluated
vasoreactivity responses following PM exposure in rats or mice. Four measured vasorelaxation induced
by endothelium-dependent and -independent agonists in the microvasculature and intrapulmonary arteries
and report impaired vasodilation following ROFA or ambient PM2 5 (SRM1648). Ti02 was reported to
have similar effects in 3 of the 4 studies (those that used the microvasculature). One study demonstrated
decreased LAV ratio in the pulmonary artery of rats following CAPs exposure and another reported
enhancement of ET-1 -induced vasoconstriction in mesenteric veins of mice with exposure to DE. The
former study findings were linked to the tracer element Si. Plasma ET-1 was increased with exposure to
gasoline emissions and appeared to be particle-independent. In contrast, ET-2 was decreased in rats
exposed to on-road highway aerosols. Increases in mRNA expression of ET-1 and ETA receptor in rat
hearts were reported following CAPs. Elevated plasma ADMA was also observed after CAPs exposure in
rats.
Blood Pressure
Epidemiologic Studies: PM2 5 was not associated with increased BP in a European multicity study
of coronary artery disease patients (Ibald-Mulli et al., 2004). Findings from single city studies are
inconsistent (Choi et al., 2007; Chuang et al., 2005b; Dales et al., 2007; Ebelt et al., 2005; Jansen et al.,
2005; Mar et al., 2005b; Zanobetti et al., 2004). Right heart pressure increases among CHF patient with
short-term PM2 5 concentration has been reported in one pilot study (Rich et al., 2008).
Human Clinical Studies: No consistent effect of exposure to fine CAPs on BP was presented in
two human clinical studies described in the 2004 PM AQCD. One new study observed no changes in BP
following exposure to fine zinc oxide (Beckett et al., 2005). However, another study did observe a
significant decrease in diastolic BP during exposure to fine CAPs relative to filtered air control (Urch et
al., 2005). Findings from new studies evaluating changes in BP following exposure to DE have been
inconsistent.
Toxicological Studies: One new CAPs study that evaluated mean BP reported a weak correlation
with accumulated PM2 5 mass and the other demonstrated elevated mean BP in SH rats during a dust
storm event.
December 2008
6-102
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Cardiac Contractility
Toxicological Studies: Only one CAPs study used an indirect measure of cardiac contractility
(QA-interval) and found increased contractility with a dust storm event.
Systemic Inflammation
Epidemiologic Studies: Most studies of inflammatory markers reviewed in the 2004 PM AQCD
focused on PM10; however, associations with BS were observed. New studies have focused on PM2 5 and
components (e.g. BC, EC, OC and sulfate) (Diez Roux et al., 2006; Dubowsky et al., 2006; O'Neill et al.,
2007; Pope et al., 2004a; Riediker et al., 2004b; Ruckerl et al., 2007b; Sullivan et al., 2007; Zeka et al.,
2006b). Although findings from from these studies were not consistent, effects of PM2.5 on inflammatory
markers were stronger with longer averaging times and among populations with preexisting diseases
(Diez Roux et al., 2006).
Human Clinical Studies: Two studies cited in the 2004 PM AQCD found no effect of exposure to
fine CAPs on markers of systemic inflammation including levels of serum amyloid A, number of
lymphocytes, or levels of IL-6 and IL-8. The 2004 PM AQCD did include one study that reported
increased peripheral blood neutrophils in subjects exposed to DE. New studies involving controlled
exposures to DE have provided very little evidence of a PM-induced increase in markers of inflammation
immediately following exposure, although one study did report significant increases in plasma levels of
IL-6 and TNF-a 24-h after exposure (Tornqvist et al., 2007).
Toxicological Studies: There were mixed results reported in the 2004 PM AQCD in regard to
systemic inflammation, with two studies that reported changes in WBC (i.e., increases in PMN and/or
decreases in lymphocytes) following CAPs or ROFA in rats. One study reported no change in systemic
inflammatory markers with CAPs. Colloidal carbon stimulated the release of PMN from bone marrow in
rabbits. Two new CAPs studies demonstrate no change or decreased WBC. One study of coal fly ash
reported elevated blood PMN and decreased lymphocytes. Fine carbon black did not induce any changes
in blood leukocytes. All of the recent studies indicate relatively little change in systemic inflammation at
16-20 h post-exposure.
Blood Coagulation
Epidemiologic Studies: Studies of markers of coagulation published since 2002 have focused on
PM2 5 and components (e.g. BC, EC OC, BC and sulfate) (Chuang et al., 2007a; Delfino et al., 2008;
O'Neill et al., 2007; Pope et al., 2004a; Sullivan et al., 2007; Zeka et al., 2006b). The most consistent
results have been observed for vWF and associations with fibrinogen are less constent. Studies of
December 2008
6-103
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
components are too few to allow conclusions to be drawn (Delfino et al., 2008; O'Neill et al., 2007)
(Chuang et al., 2007a).
Human Clinical Studies: Two studies in the 2004 PM AQCD found increased fibrinogen following
controlled exposures to fine CAPs, while another observed a decrease in factor VII levels with no change
in fibrinogen. In addition, exposure to DE was reported by one study to increase peripheral blood
platelets. One new human clinical study did not observe any change in markers of blood coagulation
following exposure to fine zinc oxide. WS has recently been shown to increase plasma factor VIII and the
factor VIII/vWF ratio in plasma (Barregard et al., 2006). New studies of exposure to DE have provided
evidence of an attenuation of t-PA release 6 h post-exposure, although no consistent diesel-induced
changes in other blood coagulation markers have been observed (Mills et al., 2005; 2007).
Toxicological Studies: A few animal toxicological studies evaluated in the 2004 PM AQCD
examined blood coagulation markers with PM2.5 exposure, with inconsistent results. One CAPs study
from Tuxedo, NY reported an increase in platelets, with another from New York City not finding any
change in blood coagulation markers. Increases in plasma fibrinogen were demonstrated for rats exposed
to ROFA or Ottawa PM (EHC-93). Of three new CAPs studies in SH rats, two reported increases in
plasma fibrinogen, while the other reported no change. Similarly, an on-road highway exposure induced
elevated plasma fibrinogen in rats. One study reported elevated von Willebrand factor in SH rats exposed
to fine CAPs. The recent studies with RBC measurements are limited to two CAPs studies and one coal
fly ash study that demonstrate mixed results.
Cardiac or Systemic Oxidative Stress
Epidemiologic Studies: New studies have observed associations of PM2.5 and components (EC,
OC, BC, vanadium and chromium) with several markers of systemic oxidative stress (Cu/Zn-SOD,
plasma proteins, 8-oxodG) (Chuang et al., 2007a; Delfino et al., 2008; Romieu et al., 2008; Sorensen et
al., 2003; 2005).
Human Clinical Studies: One study cited in the 2004 PM AQCD reported no diesel-induced
changes in plasma antioxidant concentrations or malondialdehyde. New studies have reported increases in
markers of systemic oxidative stress following controlled exposures to WS, urban traffic particles, and
DE.
Toxicological Studies: There are three new studies that report increased oxidative stress in cardiac
tissue of rats exposed to Boston CAPs; one of these also demonstrated increased levels of antioxidant
enzymes in the heart. One study found increased HO-1 mRNA expression in rat cardiac tissue following
CAPs exposure and another reported increased ROS and nitrotyrosine expression in the mouse left
December 2008
6-104
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
ventricles after exposure to carbon black. One study demonstrated oxidative stress in rat systemic
micro vasculature with ROFA instillation.
Clinical Outcomes in Epidemiologic Studies
MCAPS investigators have observed clear increases in cardiovascular admissions related to PM2 5.
CHF and IHD appear to have the strongest associations with PM2 5. Metzger et al. (2004) found positive
statistically significant associations between PM2.5 and all CVD visits, CHF visits, and peripheral vascular
visits (here defined as PVD and stroke). The French PSAS program found that PM2 5 concentration
averaged over the current and previous days was associated with increased hospitalizations for the CVD
outcomes evaluated (e.g. IHD, total CVD, cardiac disease). Although estimates from studies of
cerebrovascular diseases are less precise and consistent, ischemic diseases appear to be more strongly
associated with PM2 5 compared to hemorrhagic stroke. Too few studies have evaluated PVD to allow
conclusions to be drawn. Strong regional and seasonal heterogeneity has been observed with the strongest
estimates in the northeastern U.S. (Bell et al., 2008; Dominici et al., 2006). The null funding for PM2 5 and
out of hospital cardiac arrest and onset of MI reported by Sullivan et al. (2003; 2007) may reflect this
regional heterogeneity.
Only the SOPHIA study examined PM2 5 components and found EC and OC were associated with
cardiovascular ED visits. A handful of older publications have examined whether the associations
observed between ambient PM and CVD hospital admissions or ED visits may be attributable to particle
acidity (Burnett et al., 1997; Gwynn et al., 2000; Lippmann et al., 2000; Metzger et al., 2004).
Consistently in these studies, particle acidity has not been more strongly associated with CVD
hospitalizations or ED visits than other PM metrics.
A limited number of source apportionment studies have been conducted (Anderson and Bogdan,
2007; Sarnat et al., 2008; Schreuder et al., 2006). These studies indicate that the observed associations
between short-term increases in ambient levels of PM2 5 are largely due to traffic-related pollution and
biomass burning. However, even exposure to crustal material associated with dust storms appears to have
adverse cardiovascular health effects.
6.2.11.4. Ultrafine PM
Causal Determination
A limited number of epidemiologic studies have examined the association of ultrafine particles
with cardiovascular disease hospitalizations and ED visits. In the U.S., SOPHIA investigators in Atlanta
did not observe an association with ultrafine particles while PM2 5 associations were present. A few
studies in Europe observing associations with UFP also observed associations with PMi0. Short-term
December 2008
6-105
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
associations of UFP with subclinical markers of cardiovascular disease have been reported in a small
number of studies.
The effect of ultrafine particles on cardiovascular function has not been extensively evaluated in
human clinical studies. However, two studies have demonstrated significant decreases in HRV following
controlled human exposures to ultrafine CAPs. One of these studies also observed a significant increase in
D-dimer 18 h post-exposure in a group of healthy young adults. In addition, exposure to ultrafine EC was
recently shown to affect vasomotor function among healthy adult volunteers.
Four recent toxicological studies report cardiovascular effects with ultrafine PM exposure, although
one study used intratracheal instillation as the exposure route. The latter study reported increased infarct
size and an impaired vasodilatory response in PM-exposed mice following ischemia/reperfiision injury.
The only endpoints evaluated in multiple studies of ultrafine PM were systemic inflammation and the
results were inconsistent, perhaps due to differing blood collection times after exposure.
Based on the above findings, the evidence is inadequate to determine that a causal relationship
exists between ultrafine PM exposure and cardiovascular morbidity
Heart Rate Variability
Epidemiologic Studies: Findings from four recent studies investigating the effect of UFP on HRV
(Adar et al., 2007a; Chan et al., 2004; Park et al., 2005a; Timonen et al., 2006) were inconsistent.
Human Clinical Studies: Two new studies have demonstrated that exposure to ultrafine CAPs may
alter HRV in healthy adults and asthmatics with evidence of decreases in SDNN and LF power (Gong et
al., 2008; Samet et al., 2007). No such changes were observed following exposure to ultrafine zinc oxide
at relatively high concentrations (500 (ig/m3).
Toxicological Studies: In the only study of ultrafine CAPs, increased H and decreased SDNN
were reported only during the spring and not the summer exposure.
Arrhythmia
Epidemiologic Studies: In a recent study, a non-significant increase in ICD recorded ventricular
arrhythmias with particle number was reported in Boston (Dockery et al., 2005a; 2005b). Berger et al.
(2006) reports associations of supraventricular tachycardia and number of runs of ventricular tachycardia
with 5-day mean UFP (0.01-0.1) counts. Adverse effects on repolarization parameters were also observed
in association with UFP count (Henneberger, 2005).
Toxicological Studies: One new study exposed older rats to ultrafine carbon black and did not
report any change in arrhythmia frequency.
December 2008
6-106
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Ischemia
Epidemiologic Studies: One recent study examined UFP in relation to ST-segment depression and
found associations that were comparable to those observed with other PM size fractions (e.g. PM
PM2 5, PM10.2.5) (Pekkanen et al., 2002).
Toxicological Studies: In the first study of its kind, infarct size was nearly doubled in mice
exposed to ultrafine PM followed by ischemia/reperfusion injury to the coronary artery.
Vasomotor Function
Epidemiologic Studies: One study examining particle number count and vasomotor function
reported a nonsignificant decrease in flow mediated and nitroglycerine mediated reactivity (O'Neill et al.,
2005a).
Human Clinical Studies: Shah et al. (2008) recently demonstrated a decrease in peak forearm
blood flow during reactive hyperemia, relative to filtered air control, following exposure to ultrafine EC.
Toxicological Studies: One new study reported an attenuation of ACh-induced relaxation in aortic
rings of mice exposed to ultrafine PM that then underwent ischemia/reperfusion injury to the left anterior
descending artery. There was no difference in constriction to phenylephrine.
Blood Pressure
Epidemiologic Studies: Two recent studies examined BP in relation to PM concentration.
Increases in BP were not observed in association with UFP in a European multicity study of CAD patients
(Ibald-Mulli et al., 2004). By contrast, an increase in BP was reported in association with UFP 1-3 hours
prior to the measurement among subjects with impaired lung function (Chuang et al., 2005b).
Human Clinical Studies: Several new studies have not observed any changes in BP following
exposure to ultrafine laboratory generated surrogate particles (zinc oxide and EC).
Toxicological Studies: A new CAPs study in dogs demonstrated elevated BP that was partially
attributable to a-adrenergic receptors. The other recent CAPs study reported elevated mean BP in SH rats
during ultrafine PM exposure during spring and not summer.
Cardiac Contractility
Toxicological Studies: There is one study that used an indirect measure of cardiac contractility
(QA-interval) during ultrafine CAPs exposure and reported increased cardiac contractility during the
spring, but not summer.
December 2008
6-107
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Systemic Inflammation
Epidemiologic Studies: Previous studies included in the AQCD focused on PMi0. Asssociations
with UFP and inflammatory markers (CRP, IL-6 and sP-selectin) have been observed in new studies, but
not consistently across studies (Delfino et al., 2008; Ruckerl et al., 2007b).
Human Clinical Studies: The 2004 PM AQCD included one human clinical study that found no
effect of exposure to ultrafine EC on leukocyte activation. Three new studies have reported no changes in
markers of systemic inflammation following exposure to ultrafine particles (CAPs, zinc oxide, or EC).
However, one study observed a decrease in total leukocyte count and monocyte expression of adhesion
molecules CD54 and CD 18 in peripheral venous blood following exposure to ultrafine EC (Frampton et
al., 2006).
Toxicological Studies: There were mixed results reported in the 2004 PM AQCD in regard to
systemic inflammation, but no studies were reviewed that used ultrafine PM. One new study employed
ultrafine CAPs during a dust storm and demonstrated increased WBC even in a rodent model of
pulmonary hypertension. Two studies of ultrafine carbon black found opposite responses (elevated WBC
and lowered PMN), which may be attributable to differing post-exposure analysis times (48 and 6 h,
respectively).
Blood Coagulation
Epidemiologic Studies: Previous studies included in the AQCD focused on PM10. Asssociations
with UFP and markers of coagulation (fibrinogen, d-dimer) have been evaluated in new studies (Delfino
et al., 2008; Ruckerl et al., 2007b), but studies examining these size fractions are too few in number to
draw conclusions.
Human Clinical Studies: One human clinical study cited in the 2004 PM AQCD found no
association between exposure to ultrafine EC and markers of blood coagulation. New human clinical
studies have similarly found no pro-thrombotic effects of exposure to ultrafine EC or zinc oxide in
healthy adults or adults with coronary artery disease. However, Samet et al. (2007) did observe an
increase in concentrations of D-dimer among healthy adults following exposure to ultrafine CAPs.
Toxicological Studies: One new study of ultrafine carbon black reported increased TAT complexes
in two older rat strains. Plasma fibrinogen results were inconsistent.
Cardiac or Systemic Oxidative Stress
Epidemiologic Studies: Quasi ultrafine particles (PM < 0.25 |im) were associated with
Cu/Zn-SOD in one study (Delfino et al., 2008).
December 2008
6-108
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Clinical Outcomes in Epidemiologic Studies: Few time-series studies have compared results
using ultrafine PM. The SOPHIA study found no association between any outcome and 24-h mean levels
of UFP, but did find a number of positive associations with PM25 (Metzger et al., 2004). Andersen et al.
(2007a) found statistically significant positive associations between CVD hospitalizations and PMi0 and
PM2 5, but not with UFP. In the European HEAPSS study (Lanki et al., 2006a; von Klot et al., 2005),
results seemed qualitatively similar when comparing associations with PMi0 and UFP, but there were
insufficient data provided to standardize results to the IQR (or other distributional measure) specific to
each metric.
6.3. Respiratory Effects
6.3.1. Respiratory Symptoms and Medication Use
The 2004 PM AQCD presented evidence from epidemiologic studies of increases in respiratory
symptoms associated with PM, although this was not supported by the findings of a limited number of
human clinical studies. Recent epidemiologic studies have provided evidence of an increase in respiratory
symptoms and medication use associated with PM among asthmatic children, with less evidence of an
effect in asthmatic adults. The lack of an observed effect of PM exposure on respiratory symptoms in
human clinical studies does not necessarily contradict these findings, as no controlled human exposures to
PM have been conducted among groups of asthmatic or healthy children.
6.3.1.1. Epidemiologic Studies
The 2004 PM AQCD concluded that the effects of PMi0 on respiratory symptoms in asthmatics
tended to be positive, although they were somewhat less consistent than PMi0 effects on lung function.
Most studies showed increases in cough, phlegm, difficulty breathing, and bronchodilator use, although
these increases were generally not statistically significant for PMi0. The results from one study of
respiratory symptoms and thoracic coarse particles (Schwartz and Neas, 2000) found a statistically
significant association with cough with PM10.2.5. The results of two studies examining respiratory
symptoms and PM2 5 revealed slightly larger effects for PM2 5 than for PM10.
December 2008
6-109
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Location
Lag
Forsberg etal 1998
Mar eta! 3004
Mai nl ni 2004
Mar ef oi. 2004
Forsberg et a' 1998
Just et al. 2002
JaialuCfm et al. 2004
Aekplakorn et al 2003
Mar et al. 2004
JalakicP'". e! al. 2004
Mar ot al. 2004
Delfino fit al. 2003
Mar et al. 2004
Delfino et al 2002
Delfino et al. 2002
Delfino et a! 2002
Mortimer et al. 2002
Schilccrout et a!. 2006
Rab:novitc'n et al. 2004
Rritj-noviic*! et al. 2004
Just et al. 2002
Boezen et al 1999
Forsberg e! a! 1998
Aokptakow et al. 20G3
Mar etal. 2004
Boezen et al. 1999
Forsberg et al 1998
Aekplakorn et al. 2003
Gordian et a!. 2003
Rabinovttch el al. 2004
SchHdorcH.it ei al 2006
Slaughter et at. 2003
Forsberg et al 1998
Jalaiudln et at. 2004
Just etal. 2002
Rabinovitch et al. 2004
Slaughter et al. 2003
Sweden
"*¦*
Spokane. WA
t>poKane. WA
Sweden
Paris. Francs
Australia
Thailand
Spokane, WA
Australia
Spokane. WA
California
Spokane. WA
California
California
California
Multicity. US
US and Canada
Denver, CO
Denver. CO
Paris France
Netherlands
Sweden
Thailand
Spokane, WA
Netherlands
Sweden
Thailand
Anchorage, AK
Denver. CO
US and Canada
Seattle. WA
Sweden
Australia
Paris, France
Denver, CO
Seattle, WA
0-6
"
0
0
0-6
0-4
0-5
0
0
0
0
0
0-2
0-2
0-2
1-2
0-2
0-3
0-3
¦¦¦
¦¦¦
0-6
0
0
0-4
0-6
0
21
0-3
0-2
0
0-6
0-5
0-4
0-3
0
pivegm
-runny nose
- sputum
Mucus
Cough
Wheeze
— SOB
score >1
-any
Symptoms
¦ - 1-b max
JLiiavg
-24-h avg
current day
previous night
Resp.ratory Infection
URS
LRS
Medication Use
*
Asthma
0.5 1.0 1.5
Effect Estimates
2.0
25
Figure 6-5. Respiratory symptoms and/or medication use among asthmatic children following
acute exposure to PM10. ORs and 95% CIs standardized to increments of 10 |jg/m3.
December 2008
6-110
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Location
Lag
Rodriguez et al. 2007
Australia
0-5
Rodriguez et al. 2007
Australia
0-5
Maretal. 2004
Spokane, WA
0
Maret al. 2004
Spokane, WA
0
Barraza-Villarreal et al. 2008
Mexico City
Maret al. 2004
Spokane, WA
0
Rodriguez et al. 2007
Australia
0-5
Rodriguez et al. 2007
Australia
0-5
Aekplakorn et al. 2003
Thailand
0
Ranzi et al. 2004
Italy
0-3
Barraza-Villarreal et al. 2008
Mexico City
Maretal. 2004
Spokane, WA
0
Rodriguez et al. 2007
Australia
0-5
Rodriguez et al. 2007
Australia
0-5
Maret al. 2004
Spokane, WA
0
Maret al. 2004
Aekplakorn et al. 2003
Maretal. 2004
Aekplakorn et al. 2003
Rabinovitch et al. 2006
Slaughter et al. 2003
Slaughter et al. 2003
Spokane, WA 0
Thailand 0
Spokane, WA 0
Thailand 0
Denver, CO 1
Seattle, WA 1
Seattle, WA 1
phlegm 1-h max f
phlegm 24-h avg A
runny nose ]-
sputurn.
1-h max
24-h avg *
1-h max 4
i
24-h avg *
SOB 1-
Mucus
Cough
Wheeze
Symptoms
URS
LRS
Medication Use
Asthma
0.25
0.50
-i h
0.75 1.00
Effect Estimates
.25
1—
1.50
Figure 6-6. Respiratory symptoms and/or medication use among asthmatic children following
acute exposure to PM2.5- ORs and 95% CIs standardized to increments of 10 ng/m3.
December 2008
6-111
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Asthmatic Children
Since the 2004 PM AQCD, results have been published from several single- and multicity studies
investigating the effects of ambient PM levels on respiratory symptoms and medication use among
asthmatic children. The respiratory symptom and medication use results from these single and multicity
studies are summarized by particle size and displayed in Figures 6.5 and 6.6 and Table 6.6. A relatively
few number of studies examined these effects in healthy children, and did not identify a relationship
between ambient PM levels and respiratory symptoms or medication use. These studies are summarized
in Annex E, but will not be described in detail in this section.
Two large, longitudinal studies in urban areas of the U.S. investigated the effects of ambient PM on
respiratory symptoms and/or asthma medication use with similar analytic techniques (i.e. multistaged
modeling and generalized estimating equations [GEE]): the Childhood Asthma Management Program
(CAMP; Schildcrout et al., 2006), and the National Cooperative Inner-City Asthma Study (NCICAS;
Mortimer et al., 2002). A number of smaller panel studies conducted in the U.S. evaluated the effects of
ambient PM concentrations on respiratory symptoms and medication use among asthmatic children
(Delfino et al., 2002; 2003b; Gent et al., 2003; Rabinovitch et al., 2004; 2006; Slaughter et al., 2003).
In the CAMP study, the association between ambient air pollution and asthma exacerbations in
children (n = 990) from eight North American cities was investigated (Schildcrout et al., 2006). In
contrast to several past studies (Delfino et al., 1996; 1998), no associations were observed between PMi0
and asthma exacerbations or medication use. PM10 concentrations were measured on less than 50% of
study days in all cities except Seattle and Albuquerque. While PM10 effects were not observed for the
entire panel of children, they were observed in recent reports on the children participating at the Seattle
center (Slaughter et al., 2003; Yu et al., 2000).
The eight cities included in the NCICAS (Mortimer et al., 2002) were all in the East or Midwest:
New York City (Bronx, E. Harlem), Baltimore, Washington DC, Cleveland, Detroit, St. Louis, and
Chicago. In this study, 864 asthmatic children, aged 4-9 years, were followed daily for four, 2-week
periods over the course of nine months. Morning and evening asthma symptoms (analyzed as none vs.
any) and peak flow were recorded. For the three urban areas with data, each 10 (ig/m3 increase in the
mean of the previous 2 days (lag 1-2) PMi0, increased the risk for morning asthma symptoms (OR 1.12
[95% CI: 1.00-1.26]). This effect was robust to the inclusion of 03 (OR 1.12 [95% CI: 0.98-1.27]), though
attenuated in models including 03, S02, and N02 (OR 1.07[ 95% CI: 0.89-1.22]). In a related study,
O'Connor et al. (2008) examined the relationship between short-term fluctuations in outdoor air pollutant
concentrations and changes in pulmonary function and respiratory symptoms among children with asthma
in 7 U.S. inner-city communities. PM2 5 concentration was not significantly associated with respiratory
symptoms in this study.
December 2008
6-112
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
In a smaller panel study of asthmatic children (n = 133) enrolled in the CAMP study, daily particle
concentrations averaged over 3 central sites in Seattle was used as the exposure metric (Slaughter et al.,
2003). Children were followed for 2 months, on average. Daily health outcomes included both a
3-category measure of asthma severity based on symptom duration and frequency, and inhaled albuterol
use. In single-pollutant models, an increased risk of asthma severity was associated with a 10 (ig/m3
increase in lag 1 PM2 5 (OR 1.20 [95% CI: 1.05-1.37]) and with a 10 (ig/m3 increase in lag 0 PMi0 (OR
1.12 [95% CI: 1.05-1.22]). In copollutant models with CO, the associations remained (ORfor PM25 1.16
[95% CI: 1.03-1.30]; ORfor PM10 1.11 [95% CI: 1.03-1.19]). Associations between inhaler use and PM
was significant in single-pollutant models (RRlag 1 PM25 1.08 [95% CI: 1.01-1.15]; RRlag 0 PMi0 1.05
[95% CI: 1.00-1.09), but attenuated and no longer significant in copollutant models.
Mar et al. (2004) studied asthmatic children (n = 9) in Spokane, WA. Increases in either 0, 1 or 2
day lags of each of the PM size classes studied was associated with cough. When all lower respiratory
tract symptoms (wheeze, cough, shortness of breath, sputum production) were grouped together,
significant associations were seen for each 10 (ig/m3 increase in same-day PMi0 (OR 1.07 [95% CI:
1.00-1.14]), or lag 0 or lag 1 PM25 (OR 1.18 [95% CI: 1.00-1.38]; OR 1.21 [95% CI: 1.00-1.46],
respectively), and 10 (ig/m3 increase in lag 0 and lag 1 PML0 (OR 1.21 [95% CI: 1.01-1.44]; OR 1.25
[95% CI: 1.01-1.55], respectively). No significant effects were seen for PM10.2.5 and grouped lower
respiratory tract symptoms (2004).
Gent et al. (2003) reported on daily symptom and medication use during one summer for 271
asthmatic children living in southern New England. In single-pollutant models for users of maintenance
medication (n = 130), PM2 5 >19 (ig/m3 lagged by 1 day was associated with a 10 to 25% increase in risk
of symptoms compared to PM2 5 <6.9 (ig/m3: OR for persistent cough 1.12 (95% CI: 1.02-1.24); OR for
chest tightness 1.21 (95% CI: 1.00-1.46); ORfor shortness of breath 1.26 (95% CI: 1.02-1.54). Effects
were attenuated in models including 03 (OR for persistent cough 1.00 95% CI: 0.88-1.15]; ORfor chest
tightness 0.91 [95% CI: 0.71-1.17]; ORfor shortness ofbreath 1.20 [95% CI: 0.94-1.52]). No statistical
associations between ambient particle exposure and respiratory health were found for asthmatic children
not on maintenance medication.
December 2008
6-113
DRAFT-DO NOT QUOTE OR CITE
-------
Table 6-7. Characterization of ambient PM concentrations from studies of respiratory outcomes
and short-term exposures in asthmatic adults.
Pollutant Reference Location Mean Concentration (|jg/m3) Upper Percentile Concentrations (|jg/m3)
PM10
Boezen et al. (2005)
The Netherlands
26.6-44.1
Max: 89.9-242.2
de Hartog et al. (2003)
Multicity, Europe
19.6-36.5
Max: 67.4-112.0
Delfino et al. (2002)
Alpine, CA
20
90th: 32
Max: 42
Delfino et al. (2003a)
Los Angeles, CA
59.9
90>": 86/0/Max: 126
Delfino et al. (2004)
Alpine, CA
29.7
90th: 40.9
Max: 50.7
Delfino et al. (2006)
Southern CA
35.7-70.8
Max: 105.5-154.1
Desqueyroux et al. (2002)
Paris, France
23-28
Max: 63-84
Ebeltet al. (2005)
Vancouver, Canada
17
Max: 36
Jansen et al. (2005)
Seattle, WA
18.0
Max: 51
Maret al. (2004)
Spokane, WA
16.8-24.5
Mortimer et al. (2002)
Multicity, UC
53
Rabinovitch etal. (2004)
Denver, CO
28.1
Max: 102.0
Segala et al. (2004)
Paris, France
24.2
Max: 97.4
Schildcrout et al. (2006)
Multicity, US
17.7-32.4a
75th: 26.2-42.7
90th: 32.5-53.9
Slaughter et al. (2003)
Seattle, WA
21.0a
75th: 29.3
Steinvil et al. (2008)
Tel Aviv, Israel
64.5
75th: 60.7
von Klot et al. (2002) Erfurt, Germany 45.4 75th: 59.7
Max: 172.4
PMis
Adamkiewicz et al. (2004) Steubenville, OH 19.5 75th: 25.5
Max: 105.8
Adar et al. (2007)
St. Louis, MO
14.8-16.5
Allen et al. (2008)
Seattle, WA
11.2
de Hartog et al. (2003)
Multicity, Europe
12.8-23.4
Max: 39.8-118.1
Delfino et al. (2006)
Southern CA
3.9-6.9
Max: 8.8-11.6
DeMeo et al. (2004)
Boston, MA
10.8
Dubowsky et al. (2006)
St. Louis, MO
16
Max: 28
Ebeltet al. (2005)
Vancouver, Canada
11.4
Max: 28.7
Ferdinands et al. (2008)
Atlanta, GA
27.2
Max: 34.7
Gent et al. (2003)
CT&MA
13.1
60th: 12.1
80th: 19.0
Giradot et al. (2006)
Smoky Mountains
13.9
Max: 38.4
Jansen et al. (2005)
Seattle, WA
14.0
Max: 44
Koenig et al. (2003)
Seattle, WA
13.3
Max: 40.4
Lewis et al. (2005)
Detroit, Ml
15.7-17.5
Max: 56.1
Maret al. (2004)
Spokane, WA
8.1-11.0
Maret al. (2005)
Seattle, WA
5-26
Rabinovitch etal. (2004)
Denver, CO
10.8
Max: 53.5
December 2008
6-114
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Pollutant Reference Location Mean Concentration (|jg/m3) Upper Percentile Concentrations (|jg/m3)
Rabinovitch et al. (2006) Denver, CO TEOM: 6.5-8.2 75th: 7.9-9.9 (TEOM)
FRM: 11.2-11.8 75th: 13.3-14.1 (FRM)
Slaughter et al. (2003)
Seattle, WA
7.3a
75th: 11.3
Timonen et al. (2004)
Multicity, Europe
12.7-23.1
Max: 39.8-118.1
Trenga et al. (2006)
Seattle, WA
8.6-9.6a
75th: 13.1-14.8
Max: 40.4-41.5
Ebeltet al. (2005)
Vancouver, Canada
5.6
Max: 11.9
Maret al. (2004)
Spokane, WA
8.7-13.5
von Klot et al. (2002) Erfurt, Germany 10.3 75th: 14.6
Max: 64.3
a Median PM concentration.
Two panel studies were conducted over the course of three winters at a school in Denver
(Rabinovitch et al., 2004; 2006). In the first report, approximately 86 different children contributed data
on asthma symptoms and medication use over three consecutive winters (Rabinovitch et al., 2004). The
exposure metric was a 3-day moving average of PM25 measured at a site located next to the school for the
first 2 winters and from a central site located 4.8 km (3 miles) away for the third. A strong correlation was
observed during the first two winters between PM2 5 values measured locally and at a downtown
monitoring station (Pearson product-moment correlation = 0.93) and between PMi0 values measured
locally and at a downtown monitoring station (correlation = 0.84). Therefore, in year 3, all ambient data
were collected from nearby community monitoring stations. No significant effects were found between
asthma symptoms or medication use and PM. Rabinovitch et al. (2006) enrolled a panel of 73 children
and evaluated associations with morning maximum PM2.5 measured at the central site. PM measurements
were available hourly from 2 co-located monitors, an FRM and a TEOM monitor. Each 10 (ig/m3 increase
in morning maximum 1-h PM2 5 concentration was associated with an increased likelihood of rescue
medication use (OR for FRM exposure data 1.02 [95% CI: 1.01-1.03]; OR for TEOM 1.03 [95% CI:
1.00-1.6]). Interestingly, the association between inhaler use and particle exposure was not evident when
the 24-h average PM2 5 was used in the model.
Two smaller panel studies enrolling asthmatic children conducted by Delfino et al. (2002; 2003b)
in Southern California examined the health effects of different averaging times for PMi0 (1-h, 8-h, 24-h)
(Delfino et al., 2002), and 24-h average of two PMi0 components (EC and OC) (Delfino et al., 2003a). In
the first study, 22 children living in a "lower" pollution area were followed daily for two months in
spring. In contrast with Gent et al. (2003), significant associations with asthma symptoms (measured on a
6-point severity scale) were found only for the children not taking anti-inflammatory medication. For
these 12 children, in single-pollutant models each 10 |_ig/m3 increase in lag 0 1-h max PM10 nearly
doubled the risk of clinically meaningful symptoms (i.e., an asthma symptom score > 3) (OR 1.14 [95%
December 2008
6-115
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
CI: 1.04-1.24]) and each 10 (ig/m3 increase in 3-day average 24-h PM10 increased the risk by 1.25 (95%
CI: 1.06-1.48). No significant associations were found between exposure to ambient particles and
symptoms in the 10 children who were taking anti-inflammatory medication. No multipollutant models
were reported. The second study enrolled 22 asthmatic children living in an area of higher pollution. For
children living in this community, each 10 (ig/m3 increase in lag 0, 24-h PMi0 was associated with an
increased risk of asthma symptom score >1: OR 1.10, (95% CI: 1.03-1.19) (Delfino et al., 2003a). The
correlation among PMi0, EC and OC was substantial: 0.80 between PMi0 and either EC or OC, and 0.94
between EC and OC. Associations between EC or OC and asthma symptoms were very similar to those
for PMi0: each 3 (ig/m3 increase in lag 0, 24-h EC or 5 (ig/m3 increase in lag 0, 24-h OC was associated
with an increased risk of asthma symptoms (OR 1.85 [95% CI: 1.11-3.08] or OR 1.88 [95% CI:
1.12-3.17], respectively) (Delfino et al., 2003a).
Studies from Australia (Rodriguez et al., 2007), Europe (Ranzi et al., 2004), and Asia (Aekplakorn
et al., 2003) provide additional evidence of an association between ambient PM and respiratory symptoms
and/or medication use among asthmatic children. Two studies (Jalaludin et al., 2004; Just et al., 2002)
found no association between ambient PM levels and these health endpoints (see Figures 6-5 and 6-6).
Asthmatic Adults
Since the 2004 PM AQCD, one U.S. and several European studies have investigated the effects of
ambient PM levels on respiratory symptoms and medication use among asthmatic adults. The respiratory
symptom and medication use results from these studies are summarized by particle size and displayed in
Table 6.7 and Figure 6.7. Relatively few studies examined these effects in healthy adults, and they did not
identify a relationship between ambient PM levels and respiratory symptoms or medication use. These
studies are summarized in Annex E, but will not be described in detail in this section.
Mar et al. (2004) studied asthmatic adults (N = 16) in Spokane, WA over a 3-year time period. No
significant associations were found between PM and respiratory symptoms among the adults.
Several panel studies conducted in Europe have examined effects of daily exposures to air pollution
on adults with asthma, including studies in the Pollution Effects on Asthmatic Children in Europe
(PEACE) study (2005), Exposure and Risk Assessment for Fine and Ultrafine Particles in Ambient Air
(ULTRA) study (de Hartog et al., 2003), in Germany (von Klot et al., 2002), and in Paris (2002a; 2004).
Boezen et al. (2005) enrolled 327 elderly adults in the Netherlands to examine the role of AHR and IgE
levels in susceptibility to air pollution. For subjects with both AHR (defined as > 20% FEVi decline at <
2 mg cumulative methacholine) and high total IgE (>20 kU/L), each 10 (ig/m3 increase in lag 2 PMi0
concentration was associated with an increased risk of upper respiratory symptoms (URS) among males
(OR 1.06 [95% CI: 1.02-1.10]), and at lag 0 with increased cough among females (OR 1.04 [95% CI:
December 2008
6-116
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
1.00-1.08]). Each 10 (ig/m3 increase in BS at lag 0, lag 1, and the 5-day mean was associated with URS
and cough among males. The strongest association in both cases was for the 5-day mean (OR for URS
1.43 [95% CI: 1.20-1.69]; OR for cough 1.16 [95% CI: 1.05-1.29]). The authors suggest that the sex
differences observed may be expolained by differential daily exposure to traffic exhaust experienced by
men compared to women(Boezen et al., 2005)
As part of the multicenter ULTRA study, de Hartog et al. (2003) enrolled 131 older adults with
coronary artery disease in three cities (Amsterdam, Erfurt [Germany], and Helsinki). Pooling data from
all three cities, significant associations were observed between PM2 5 and shortness of breath and phlegm:
each 10 |ig/nr' increase in the 5-day average PM2 5 was associated with an increased risk of symptoms
(OR for shortness of breath 1.12 [95%CI: 1.02-1.24]; ORfor phlegm 1.16 [95% CI: 1.03-1.32]). Unlike
fine particles, ultrafine particles were not consistently associated with symptoms.
In a study that took place in Erfurt, Germany, von Klot et al. (2002) examined daily, winter time
exposure to ambient PM10-2.5, PM2.5-0.01 and PM0.1-0.01 particles and respiratory health effects in 53 adult
asthmatics. The authors examined associations between wheeze, use of inhaled short-acting p2-agonists or
inhaled corticosteroids and exposure to particles in single and multipollutant models. Particle exposure
metrics examined included same-day, 5-day and 14-day averages. No significant effects were observed
for wheeze and exposure to PM10_2.5 or PM2 5.0 01 for any averaging time. The strongest association
between wheeze and exposure to ultrafine particles was for a 14-day average: each 7,700 increase in the
NCo.01-0.1 increased the risk of wheeze by 27% (OR 1.27 [95% CI: 1.13-1.43]). The effect was attenuated
in copollutant models that also included PM2 5-0 01 (OR 1.12 [95% CI: 1.01-1.24]), N02 (OR 1.12 [95% CI:
0.99-1.26]), CO (OR 1.05 [95% CI: 0.92-1.19]) or S02 (OR 1.14 [95% CI: 1.04-1.26]). The correlations
between ultrafine particles and two gaseous pollutants, N02 and CO, were high: 0.66 for each.
In the same study, no association was found between exposure to thoracic coarse, fine or ultrafine
particles and use of short-acting inhalers, though there was an association with maintenance medication.
Increased likelihood of maintenance medication was significantly associated with PM of all sizes and all
averaging times (same day, 5- and 14-day averages) and gaseous copollutants in single or multipollutant
models. The strongest effects were seen for 14-day averages of PMi0.2.5 (for each 10 (ig/m3 increase OR
1.43 [95% CI: 1.28-1.60]), PM2.5_0.oi (for each 20 (ig/m3 increase OR 1.54 [95% CI: 1.43-1.66]), NC0 0i-o 1
(for each 7,700 increase OR 1.45 [95% CI: 1.29-1.63]). For PM2 5.0 oi, effects were unchanged in
copollutant models, including a model with ultrafine particles. The authors conclude that this is evidence
for independent effects of fine and ultrafine particles (von Klot et al., 2002).
December 2008
6-117
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Location
Desqueyroux etal. 2002 Paris. France
¦ PMI0
APM.a-a
* PM,«
Lag
3-5
Mar et a!. 2004
Spokane, WA
0
Boezen et al. 2005
Netherlands
0-4
Hi Hermann etal. 1998
Netherlands
0
Segala et al. 2004
Paris, France
0
Mar et al. 2004
Spokane, WA
0
Mar et al. 2004
Spokane, WA
0
Mar et al. 2004
Spokane, WA
0
Mar et al. 2004
Spokane, WA
0
Mar et al. 2004
Spokane, WA
0
Hiltermann etal. 1998
Netherlands
0
von Klot et al. 2002
Germany
0-4
Mar et al. 2004
Spokane, WA
0
Mar et al. 2004
Spokane, WA
0
Mar et al. 2004
Spokane, WA
0
Mar et al. 2004
Spokane, WA
0
de Hartog et al. 2003
Netherlands
0-4
Mar et al. 2004
Spokane, WA
0
Mar et al. 2004
Spokane, WA
0
Mar et al. 2004
Spokane, WA
0
Johnston et al. 2006
Australia
0-5
Mar et al. 2004
Spokane, WA
0
Mar et al 2004
Spokane, WA
0
Johnston et al 2006
Australia
0-5
Boezen et al. 2005
Netherlands
0-4
Mar et al. 2004
Spokane, WA
0
Mar et al. 2004
Spokane, WA
0
Hiltermann et al. 1998
Netherlands
0
Mar et al. 2004
Spokane, WA
0
von Klot et al. 2002
Germany
0-4
Mar et al. 2004
Spokane, WA
0
Mar et al. 2004
Spokane, WA
0
Mar et al. 2004
Spokane, WA
0
de Hartog et al. 2003
Netherlands
0-4
Asthma
Cough
Runny Nose
Sputum
LRS
Medication
Use
Mucus
Runny Nose
Sputum
*
Phlegm
. Runny Nose
_Sputum
Asthma
Sym ptoms
URS
Wheeze
. SOB (Shortness of Breath)
-SOB
.SOB
SOB
0.50
—1 1 1—
0.75 1.00 1.25
Relative Risk or Odds Ratio
1.50
Figure 6-7.
Respiratory symptoms and/or medication use among asthmatic adults following acute
exposure to particles. Summary of studies using 24-h averages of PM10, PM2.5,
PM10-2.5. ORs and 95% CIs were standardized to increments of 10 jjg/m3.
In Paris, Segala et al. (2004) recruited 78 adults from an otolaryngology clinic and followed them
for three months. Both PM10 and BS (which were very highly correlated |r =.881) were associated with
December 2008
6-118
DRAFT-DO NOT QUOTE OR CITE
-------
1 cough: OR 1.24 (95% CI: 1.01-1.52) for a 10 (ig/m3 increase in mean 0-4 day PM10 and OR 1.18 (95%
2 CI: 1.02-1.39) for a 10 (ig/m3 increase in BS.
Reference Outcome Pollutant
, ChilHrpn
•
Slaughter etal. (2003) asthma severity PM10
PM10 + CO
Mortimer et al. (2002) am asthma symptoms PM10
PM10 + 03
Aekplakornetal. (2003) cough PM10
PM|o + SO2
Desqueyrouxetal. (2002b) asthma attack PM10
PM10+ NO2
PM10 + O3
PM10 + S02
# Adults
Slaughter etal. (2003) asthma severity PM25
PM2 5 + CO
Aekplakorn etal. (2003) cough PM25
pm2.5+so2
, Children
Aekplakorn etal. (2003) cough PM10.2.5
PM,o,.5+S02
, Children
1
1 1 1 1
0.7 0.9 1.1 1.3 1.5 1.7 1.9
Effect Estimates
Figure 6-8 Respiratory symptoms following acute exposure to particles and additional criteria
pollutants.
December 2008
6-119
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Also in Paris, 60 severe asthmatics were followed for 13 months and the relationship between daily
air quality (including 24-h PM10 as measured at the site nearest to the subject's home) and asthma attack
(defined as the need to increase rescue medication use and one or more positive signs on clinical exam of
wheezing, expiratory brake, thoracic distention, hypertension with tachycardia, polypnea) were examined
with GEE models (Desqueyroux et al., 2002a). Each 10 (ig/m3 increase in PMi0 increased the risk of
asthma attack, but only after lags of 3 to 5 days. The strongest effect was seen for the mean lag of days 3
to 5 (OR 1.21 [95% CI: 1.04-1.40]). Effect sizes were larger among patients not on regular oral steroid
therapy: for PMi0 lag 3-5 (OR 1.41 [95% CI: 1.15-1.73]). This effect persisted in multipollutant models
for wintertime levels ofPM10 and S02 (OR 1.51 [95% CI: 1.20-1.90]) orN02 (OR 1.43 [95% CI:
1.16-1.76]), but not in summertime models with ozone (OR 1.09 [95% CI: 0.71-1.67]).
Copollutant Models
A limited number of respiratory symptoms studies reported results of copollutant models.
Generally, the associations between respiratory symptoms and PM were robust to the inclusion of
copollutants (Figure 6-8), though Desqueyroux et al. (2002a) indicate the effects of PM may be
potentiated by N02 and S02 during the winter months. Gent et al. (2003) also reported the results of
copollutant models, though the categorical exposure groups used in the analysis did not allow these
results to be included in Figure 6-8. As reported above, the investigators found that effects were
attenuated in models including 03.
6.3.1.2. Human Clinical Studies
Neither new controlled human exposure studies, nor studies cited in the 2004 PM AQCD have
found significant effects of CAPs on respiratory symptoms among healthy or asthmatic adults, or among
older adults with COPD (Gong et al., 2000; 2003a; 2004a; 2004b; Petrovic et al., 2000). One new study
reported an increase in respiratory symptoms (upper and lower airways) among healthy volunteers (19-59
years old) during a 2-h exposure to road tunnel traffic (Larsson et al., 2007). However, information on
specific respiratory symptoms (e.g., throat irritation, wheeze or chest tightness) is not provided. In
addition, this study only evaluated respiratory symptoms pre- versus post-exposure, and did not compare
response with a filtered air control exposure. Pietropaoli et al. (2004) found no association between
exposure to ultrafine carbon particles and respiratory symptoms in healthy adults at concentrations
between 10 and 50 (ig/m3, or asthmatics at a concentration of 10 |ig/nr\ Beckett et al. (2005) exposed
healthy subjects to ultrafine and fine zinc oxide (500 |ig/m3) and observed no difference in respiratory
symptoms compared to filtered air control 24-h following exposure. These new studies confirm previous
reports that have found no association between PM exposure and respiratory symptoms.
December 2008
6-120
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
6.3.2. Pulmonary Function
Epidemiologic studies cited in the 2004 PM AQCD observed small decrements in pulmonary
function associated with both PMi0 and PM25. The majority of human clinical studies reported no effect
of PM on pulmonary function, while the results from toxicological studies were mixed, with some
evidence of changes in tidal volume and respiratory rate following exposure to CAPs. Epidemiologic
studies published since the 2004 PM AQCD have reported an association between PM2 5 concentration
and decrements in FEVi, particularly among asthmatic children. These findings are coherent with results
from a number of recent toxicological studies which have observed increases in airways
hyperresponsiveness following CAPs exposure. Results from recent human clinical studies have been
inconsistent, with some studies demonstrating small decreases in arterial oxygen saturation or maximal
mid-expiratory flow following exposure to CAPs or elemental carbon. It is interesting to note that these
effects appear to be more pronounced among healthy adults than adults with asthma or COPD.
6.3.2.1. Epidemiologic Studies
The 2004 PM AQCD concluded that both PMi0 and PM2 5 appeared to affect lung function in
asthmatics. Ultrafine particles did not appear to have any notably stronger effect than other
larger-diameter fine particles. Few analyses were able to clearly distinguish the effects of PM10 and PM2 5
from other pollutants. Results for PM10 peak flow analyses in non-asthmatic studies were inconsistent,
with fewer studies reporting results statistically significant association.
Asthmatic Children
Seven recent panel studies have been conducted in the U.S. examining the association of exposure
to ambient PM and lung function in asthmatic children (2003a in Southern California; 2004; Dusek et al.,
2006 and Allen et al., 2008 in Seattle; 2005 in Detroit; 2004 in Denver);0'Connor, 2008. Mean
concentration data from these studies are summarized in Figure 6-8.
In the Inner-City Asthma Study (ICAS), FEBi and PEFT were significantly related to the 5-day
average of PM2 5, but not to the 1-day average concentration (O'Connor et al., 2008). The risk of
experiencing a percent-predicted FEVi more than 10% below personal best was significantly related to
the 5-day average concentration of PM2 5 (1.14 [95% CI: 1.01-1.29]). The risk of experiencing a percent-
predicted PEFRmore than 10% below personal best was significantly related to PM2 5 (1.18 [95% CI:
1.03-1.35]). This effect remained statistically significant in multipollutant models with 03 and N02 for the
FEV1 effect, but not the PEFR effect.
December 2008
6-121
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
The Denver study (Rabinovitch et al., 2004), described in Section 6.3.1.1 also examined daily
forced expiratory volume in 1 sec (FEV,) and peak expiratory flow (PEF) in 86 asthmatic children over
the course of 3 winters (some subjects participated in more than one winter). Lung function measurements
were performed under supervision daily at the elementary school where all subjects attended, and without
supervision every evening and on nonschool days. As described above, the authors chose to use a 3-day
moving average of 24-h PM25 or PMi0 as the exposure metric. No statistical associations were observed
between morning or afternoon FEVi or PEF and particle exposure. The same group of researchers used
regression calibration to estimate personal exposures to ambient PM2 5 and found that a 10 (ig/m3 increase
in PM2 5 was associated with a 2.2% (95% CI: 0.0-4.3) decrease in FEVi at a 1-day lag as compared with
the estimate of a 1.0% decrease in FEVi using ambient PM25 concentrations from fixed monitors. These
results underscore the effects of exposure error on epidemiologic study results; the effect estimate using
an estimate of personal exposure to ambient PM2 5 was twice that for central site PM2 5.
From winter 2001 to the spring of 2002, the same number (n = 86) of primary school-age asthmatic
children participated in six, 2-wk seasonal assessments of lung function in Detroit (Lewis et al., 2005).
Using a protocol similar to that used in Rabinovitch et al. (2004), morning lung function measurements
(FEVi, PEF) were self-administered at school under supervision by research staff. Evening and weekend
measurements were made without supervision by research staff at home. Community-level exposure was
assessed using monitors placed on a school roof top of both of the 2 communities. Most of the subjects
(82 of 86) lived within 5 km of their respective community monitors. In single-pollutant models using
GEE and only among children reporting the use of maintenance medication (corticosteroids), each
10 (ig/m3 increase in lag 2 PMi0 was associated with a decrease in the lowest daily percent predicted
FEVi (a reduction of 1.15%, [95% CI: -2.1 to -0.25]). Among children reporting presence of URI on the
day of lung function measurement, increases in the average of lag 3-5 of either PM2 5 or PMi0 resulted in a
decrease in the lowest daily FEVi (for a 10 (ig/m3 increase in PM2 5 the reduction was 2.24% [95% CI:
-4.4 to -0.25]; and for a 10 (ig/m3 increase in PMi0 the reduction was 2.4% [95% CI: -4.5 to -0.3]). In
copollutant models that included one particle pollutant and ozone, and among children using maintenance
medication, lag 3-5 PM2 5 continued to be associated with lowest daily FEVi as well as diurnal FEVi
variability: each 10 (ig/m3 increase was associated with a 2.23% decrease in FEVi (95% CI: -3.92 to
-0.57) and a 2.22% increase in FEVi variability (95% CI: 1.0 to 3.50). Increases in lag 1 or lag 2 of PMi0
were also significantly associated with FEVi and FEVi diurnal variability in copollutant models. The
strongest association was with lag 2 for diurnal variability (for each 10 (ig/m3 increase variability
increased by 7.0% [95% CI: 4.2 to 9.6). It is unclear what role the lack of supervision during the evening
and weekend measures may have had on these diurnal results.
December 2008
6-122
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Two panel studies in Southern California examined the association of PM exposure on lung
function in asthmatic children (Delfino et al., 2003a; 2004). In Delfino et al. (2003a), described above, no
association between exposure to particles and PEF was found for 22 Hispanic, asthmatic children living in
an area of relatively high pollution. In Delfino et al. (2004) 19 asthmatic children aged 9-17 were
followed for 2 weeks and daily, self-administered FEVi measurements were taken. Particle exposures
studied included central-site PMi0 in addition to personal PM (in the range of 0.1-10 (.un range, with the
highest response in the fine PM range), and home stationary measurements of both PMi0 and PM2 5. The
authors found significant inverse associations between percent expected FEVi and PM indicators. The
strongest association for exposure to personal PM was for a 5-day moving average of 12-h daytime PM:
for each 10 (ig/m3 increase, FEVi decreased by 7.1% (95% CI: -9.9 to -2.9). Effects for all stationary sites
(inside and outside of residence, central site) for PM2 5 were on the order of 1 to 2% reductions in FEVi,
with the strongest associations for the 5-day moving average (given in figures only). Likewise for PMi0
measured at stationary sites, the strongest effects were for 5-day moving averages and ranged from
approximately 3.8% reduction associated with indoor monitors to about 1.5% for both the outdoor and
central site monitors (given in figures only). A helpful comparison among all 24-h measures is given for
10 |_ig/m3 increases in personal PM and PM2 5 associated with decreases in percent predicted FEVi: an
increase of 10 (ig/m3 personal PM is associated with a decrease in FEVi of 3.0% (95% CI: -5.6 to -0.5);
10 (ig/m3 increase in indoor PM with 2.4% decrease (95% CI: -4.2 to -0.6); 10 (ig/m3 increase in outdoor
PM with 1.5% decrease (95% CI: -3.4, 0.1); 10 (ig/m3 increase in central site PM with 0.9% decrease
(95% CI: -2.6, 0.5).
Trenga et al. (2006) also reported associations among personal, residential, and central site PM2 5
and lung function in 17 asthmatic children in Seattle. The only significant association with decline in
FEVi was with indoor measurements of PM2 5: each 10 |_ig/m3 increase in lag 1 indoor PM2 5 was
associated with a decline in FEVi of 64.8 ml (95% CI: -111.3 to 18.3) (a 3.4% decline from the mean of
1.9 L). Indoor PM2 5 (lag 1) was also associated with declines in PEF (by 9.2 L/min [95% CI: -17.5 to
-0.9], a 3.6% decline from the 254 L/min average) and in MMEF for the 6 subjects not taking
anti-inflammatory medication (by 12.6 L/min [95% CI: -20.7 to -4.6], a 13.7% decline from the 92 L/min
average). Personal PM2 5 (lag 1) was only statistically associated with PEF for the 6 subjects not on
anti-inflammatory medication: each 10 (ig/m3 increase resulted in a 10.5 L/min ([95% CI: -18.7 to -2.3], a
4.5% decline from the 233 L/min average) reduction in PEF. Anti-inflammatory medication use
significantly attenuated associations with PM25. No statistically significant association of the ambient
PM25 mixture.
Also in Seattle, Allen et al. (2008) evaluated the effect of different PM2 5 exposure metrics in
relation to lung function among children in woodsmoke-impacted areas. The authors found that the
December 2008
6-123
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
ambient-generated component of PM2 5 exposure was associated with decrements in lung function only
among children not using inhaled corticosteroids, whereas no association ewas reported with the
nonambient exposure component. All of the ambient concentrations were associated with decrements in
both PEF and MEF. There were no associations between any exposure metrics and FVC. The authors
suggest that lung function may be especially sensitive to the combustion-generated component of ambient
PM2 5, whereas airway inflammation may be more closely related to some other consitutent
Moshammer and Neuberger (2003) used a novel technique for assessing exposure to PM in a study
they conducted in Austria. They employed a diffusion charging particle sensor (model LQ 1-DC, Matter
Engineering AG, Wohlen, Switzerland) and a photoelectric aerosol sensor (model PAS 2000 CE,
EcoChem Analytics, League City, TX) to relate the spirometry scores of Upper Austrian children
aged 7-10 to particle surface area and particle-bound PAH concentration, respectively. Details on these
methods for measuring surface area and PAH can be found in Shi et al. (2001) and Burtscher (2005),
respectively. By measuring the surface area distribution, it was possible to understand potential for
contact area with the respiratory tract cells. The authors found that acute decrements of pulmonary
function (FVC, FEVi, MEF50) were related to the active surface of particles after adjustment for PMi0.
For short-term lung impairments, this indicates that active particle surface is a better index of exposure
than PM mass.
A number of additional panel studies conducted outside of the U.S. also examined lung function
using more traditional exposure metrics. Several European and Asian studies reported associations with
decrements in pulmonary function (FEVi, FVC, FEF, MEF, PEFR) (Hogervorst et al., 2006; Hong et al.,
2007; Moshammer et al., 2006; Peacock et al., 2003; Peled et al., 2005). Others found little evidence for a
relationship between PM and daily changes in PEF after correction for the confounding effects of
weather, trends in the data, and autocorrelation (Fischer et al., 2002; Holguin et al., 2007; Janssen et al.,
2003; Just et al., 2002; Preutthipan et al., 2004; Ranzi et al., 2004; Ward, 2003).
Adults
Trenga et al. (2006) examined personal, residential, and central site monitoring of particles and the
relationship with lung function in Seattle. In models controlling for gaseous copollutants (CO, N02),
adults, regardless of COPD status, experienced a significant decline in FEVi associated only with
measurements of PM25 at the central site: each 10 |ig/m3 increase in lag 0 PM2 5 was associated with a
35.3 ml (95% CI: -70 to -1.0) decrease in FEVi. This represents a 2.2% decline in mean FEVi (mean 1.6
L during the study). Results for personal, indoor and outdoor measures of PM2 5 were inconsistent. No
significant associations were reported with outdoor PMio_2 5.
December 2008
6-124
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Giradot et al. (2006) assessed the effects of PM25 on the pulmonary function of adult day hikers in
the Great Smoky Mountains National Park. Hikers performed spirometry both before their hike and when
they returned from their hike. The authors reported no statistically significant responses in pulmonary
function with an average of 5 h of outdoor exercise at ambient PM2 5 levels that were below the current
NAAQS. Specifically, posthike percentage changes in FVC, FEVi, FEVi/FVC, FEF25.75, and PEF were
not associated with PM2 5 exposure.
Ebelt et al. (2005) developed an approach to separately estimate exposures to PM of ambient and
nonambient origin based on a mass balance model. These exposures were linked with respiratory and
cardiovascular health endpoints for 16 patients with COPD in Vancouver, Canada (mean age 74 years).
Effect estimates for estimated ambient exposure were generally equal to or larger than those for the
respective ambient concentration levels for post-FEV and AFEVi, and were statistically significant for all
AFEVi comparisons (estimated from figure).
Several studies outside of the U.S. and Canada examined the relationship between PM
concentrations and lung function and all reported a decrease in lung function in adults (FEVi, FVC,
PEFR) associated with PM exposure (Boezen et al., 2005; Bourotte et al., 2007; Lagorio et al., 2006; Lee
et al., 2007; McCreanor et al., 2007; Penttinen et al., 2006).
6.3.2.2. Human Clinical Studies
As with respiratory symptoms, there is very little evidence from human clinical studies of
PM-induced changes in pulmonary function. One study cited in the 2004 PM AQCD noted a significant
decrement in thoracic gas volume in healthy adults following a 2 hour exposure to PM2 5 CAPs
(92 |ig/m3): however, no significant changes were observed in spirometry, diffusing capacity (DLCO),
total lung capacity, or airway resistance (Petrovic et al., 2000). Other studies have found no significant
changes in pulmonary function in healthy adults following exposure to inhaled iron oxide particles (Lay
et al., 2001), or in healthy and asthmatic adults following exposure to CAPs (Ghio et al., 2000; Gong et
al., 2000; 2003a; 2004b). While Gong et al. (2004a) did not observe a significant association between
exposure to PM2 5 CAPs and spirometry in older subjects (60-80 years old), the investigators did report a
decrease in oxygen saturation immediately following CAPs exposure. This effect was observed more
consistently in healthy older adults than in older adults with COPD. These findings were confirmed by a
subsequent study conducted by the same laboratory (Gong et al., 2005). The authors also observed a small
decrease in maximal mid-expiratory flow (MMEF) following a 2-h exposure to PM2 5 CAPs (200 (.ig/nr1)
which was more pronounced in healthy subjects. Pietropaoli et al. (2004) observed a significant reduction
in MMEF and DLCO in healthy adults 21 hours after a 2-h exposure to ultrafine carbon particles
(50 |ig/m3). This reduction in DLCO may reflect a PM-induced vasoconstrictive effect on the pulmonary
December 2008
6-125
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
vasculature. Among a group of healthy and asthmatic adults exposed to ultrafine particles (Los Angeles,
mean concentration 100 |ig/m3). Gong et al. (2008) observed small, yet statistically significant decrements
in arterial oxygen saturation immediately following exposure, 4 h post-exposure, and 22-h post-exposure
(0.5% mean decrease relative to filtered air across all time points, p < 0.05). A statistically significant
decrease in FEVi was also observed, but only at 22-h post-exposure (2% decrease relative to filtered air, p
< 0.05). The responses demonstrated in this study were not significantly affected by health status.
Conversely, Samet et al. (2007) did not observe any significant changes in pulmonary function in healthy
adults following exposures to ultrafine, fine, and thoracic coarse fraction CAPs generated from ambient
air in Chapel Hill, NC.
Taken together, the majority of human clinical studies do not provide evidence of PM-induced
changes in pulmonary function; however, some investigators have observed decrease in DLCO
(Pietropaoli et al., 2004), MMEF (Gong et al., 2005; Pietropaoli et al., 2004), and oxygen saturation
(Gong et al., 2004a; 2005; 2008) following exposure to PM.
6.3.2.3. Toxicological Studies
The 2004 PM AQCD included three animal toxicological studies which measured pulmonary
function following multi-day short-terminhalation exposure to CAPs. A decreased respiratory rate was
noted in one study involving dogs. Increased tidal volume was observed in one study involving rats while
no changes were observed in the other rat study. Airway hyperresponsiveness (AHR) was found in 4
studies of mice, healthy rats or spontaneously hypertensive (SH) rats exposed to ROFA by intratracheal
instillation or inhalation. Studies conducted since the last review are discussed below.
Spontaneously hypertensive Wistar Kyoto rats (SH rats) exposed to Tuxedo, NY CAPs via
nose-only inhalation for 4 h (mean concentration 73 (ig/m3; single-day concentrations 80 and 66 (ig/m3;
2/2001 and 5/2001 respectively) had a statistically significant decreased respiratory rate compared with
air-exposed controls (Nadziejko et al., 2002). This measure was obtained from BP fluctuations using
radiotelemetry. The decrease in respiratory rate of 25-30 breaths/min was an immediate response to
CAPs, beginning shortly after the exposure began and ceasing with the end of exposure. It was
accompanied by a decrease in HR (see Section 6.2.1.3). Rats were also exposed to fine (MMAD 160 nm;
49-299 |ig/m3) and ultrafine sulfuric acid (MMAD 50-75 nm; 140-750 |ig/m3) (Nadziejko et al., 2002);
because sulfuric acid aerosols have the potential to activate irritant receptors. Irritant receptors, found at
all levels of the respiratory tract, include rapidly-adapting receptors and sensory C-fiber receptors
(Coleridge and Coleridge, 1994; Widdicombe and Lee, 2001; 2006; 2003). Activation of these vagal
afferents causes central nervous system reflexes resulting in bronchoconstriction, mucus secretion,
mucosal vasodilation, cough, and apnea followed by rapid shallow breathing. Besides effects on the
December 2008
6-126
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
respiratory system, effects on the cardiovascular system can also occur including bradycardia and
hypotension or hypertension. Fine sulfuric acid induced an overall decrease in respiratory rate, with
ultrafine sulfuric acid resulting in elevated respiratory rate compared to control (Nadziejko et al., 2002).
The authors suggested that both CAPs and fine sulfuric acid aerosols activated sensory irritant receptors
in the airways, resulting in a decreased respiratory rate. The response to ultrafine sulfuric acid aerosols
differed from the other responses and was thought to be due to deposition of ultrafine particles deeper into
the lung with the subsequent activation of pulmonary irritant receptors which trigger an increase in
respiratory rate. Since lung irritant receptors in both airways and pulmonary region act via
vagally-mediated parasympathetic pathways, this study indicates a role for neural reflexes in respiratory
responses to CAPs.
McQueen et al. (2007) also investigated the role of vagally-mediated pathways in respiratory
responses to PM. Respiratory min volume (RMV) was increased in anesthetized Wistar rats 6 h after
treatment with 500 |_ig DEP (SRM2975) by intratracheal instillation. This response was blocked by
sectioning the vagus nerves or pretreatment with atropine. The absence of respiratory response with
vagotomy or atropine indicated that the increase in RMV following DEP exposure involved a neural
reflex acting via vagal afferents. No statistically significant changes in mean BP, HR or HRV were
observed in response to DEP in this study. Vagally-mediated inflammatory responses to DEP were also
observed in this study and are discussed in Section 6.3.3.3.
Kodavanti et al. (2005) measured respiratory frequency 1 day after a 2-day exposure of SH and
WKY rats to CAPs from RTP, NC (mean mass concentration range 144-2,758 (ig/m3; less than 2.5 um in
size; 8/27-10/24/2001) for 4 h/day. Increases in inspiratory and expiratory times were seen in SH, but not
WKY rats exposed to CAPs compared with filtered air controls.
Effects of CAPs on pulmonary function were also investigated in a rat model of pulmonary
hypertension using Sprague Dawley rats pre-treated with monocrotaline (Lei et al., 2004b). In this study,
rats were exposed to CAPs from an urban high traffic area in Taiwan (mean mass concentration
371 (.ig/m') for 6 h/day on 3 consecutive days and pulmonary function was evaluated 5 hours
post-exposure using whole-body plethysmography. A statistically significant decrease in respiratory
frequency and increase in tidal volume were observed following CAPs exposure, along with an increase
in airway responsiveness (measured by Penh) following methacholine challenge.
Li et al. (2007) exposed BALB/c and C57BL/6 mice to clean air or to low dose DE (containing
100 (ig/m3 DEP) for 7 h/day and 5 days/week for 1, 4 and 8 weeks. Average gas concentrations were
reported to be 3.5 ppm CO, 2.2 ppm N02, and less than 0.01 ppm S02. Airways hyperresponsivenesss
was evaluated by whole body plethysmography at day 0 and after 1, 4 and 8 weeks of exposure. Exposure
to DE for 1 week resulted in an increased sensitivity of airways to methacholine measured as Penh, in
December 2008
6-127
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
C57BL/6 but not BALB/c mice. Other responses of this study are discussed in Sections 6.3.3.3. and
6.3.4.2.
In a study by Last et al. (2004), BALB/c mice were exposed to 250 (ig/m3 laboratory-generated
iron-soot (size range 80-110 nm; about 200 (ig/m3 as soot) for 4 h/day and 3 days/week for 2 weeks. Lung
function was measured as Penh by whole-body plethysmography after challenge with methacholine. No
increase in airway responsiveness was observed following 2 week exposure to iron-soot. Other findings
of this study are reported in Sections 6.3.3.3, 6.3.5.2 and 7.3.2.2.
In summary, several recent studies demonstrated alterations in respiratory frequency and AHR
following short-term exposure to CAPs and DE. Two studies provide evidence for the involvement of
irritant receptors and vagally-mediated neural reflexes in mediating changes in respiratory functions.
6.3.3. Pulmonary Inflammation
The discussion of the effects of PM on pulmonary inflammation in the 2004 PM AQCD was
limited by a relative lack of information from human clinical and toxicological studies. Although no
epidemiologic studies of pulmonary inflammation were described in the 2004 PM AQCD, several recent
studies have observed a positive association between PM concentration and exhaled nitric oxide. New
human clinical and toxicological studies have also generally observed an increase in markers of
inflammation in the airways following exposure to PM.
6.3.3.1. Epidemiologic Studies
No epidemiologic studies of pulmonary inflammation were described in the 2004 PM AQCD.
Exhaled Nitrogen Oxide - Asthmatic Children
Exhaled NO, a biomarker for airway inflammation, was the outcome studied in panels of asthmatic
children in Southern California (Delfino et al., 2006) and Seattle (2008; Koenig et al., 2003; Koenig et al.,
2005; Mar et al., 2005a). Mean concentration data from these studies are summarized in Table 6-7.
Delfino et al. (2006) followed 45 asthmatic children for 10 days with offline fractional eNO and
examined the associations with exposures to personal PM2 5 and 24-h PM2 5, EC and OC as well as
ambient PM2 5, EC and OC. The strongest associations were between eNO and 2-day average pollutant
concentrations: for a 10 (ig/m3 increase in personal PM2 5, eNO increased by 0.46 ppb (95% CI:
0.04-0.79); for 0.6 (ig/m3 personal EC, eNO increased by 0.7 ppb (95% CI: 0.3-1.1). An association with
exposure to ambient PM2 5 was only statistically significant in 19 subjects taking inhaled corticosteroids:
for each 10 (ig/m3 increase in PM2 5, eNO increased by 0.77 ppb (95% CI: 0.07-1.47).
December 2008
6-128
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
In a panel of 19 asthmatic children in Seattle, effects were observed only among the 10 non-users
of inhaled corticosteroids. For each 10 (ig/m3 increase in personal, outdoor, indoor, or central site PM2 5,
eNO increased from 3.82 ppb (associated with central site, 95% CI: 1.22-6.43) to 4.48 ppb (with personal
PM2 5, 95% CI: 1.02-7.93) (Koenig et al., 2003). Further analysis examining the association between eNO
and outdoor and indoor-generated particles suggested that eNO was associated more strongly with
ambient particles, but only for non-users of medication: each 10 (ig/m3 increase in estimated ambient
PM2 5 results in an increase in eNO of 4.98 ppb (95% CI: 0.28-9.69) (Koenig et al., 2005).
Also in Seattle, WA, Mar et al. (2005a) examined the association between eNO and ambient PM2 5
concentration among children (aged 6-13 years) recruited from an asthma/allergy clinic. FeNo was
associated with hourly averages of PM2 5 up to 10-12 hours after exposure. Each 10 (ig/m3 increase in
1-hour mean PM2 5 concentration was associated with a 6.99 ppb increase in eNO (95% CI: 3.43-10.55)
among children not taking inhaled corticosteroids, but associated with only a 0.77 ppb decrease in eNO
(95% CI: -4.58to3.04) among those taking inhaled corticosteroids.
Allen et al. (2008) evaluated the effect of different PM2 5 exposure metrics in relation to airway
inflammation among children in woodsmoke-impacted areas of Seattle. The authors found that the
ambient-generated component of PM2 5 exposure was associated with respiratory responses, both airway
inflammation and decrements in lung function, whereas the nonambient PM2 5 exposure component was
not. They did note, however, different relationships for airway inflammation and decrements in lung
function, with the former significantly associated with total personal PM2 5, personal light-absorbing
carbon (LAC), and ambient generated personal PM2 5 and the latter related to ambient PM2 5 and its
combustion markers. The different results between forced exhaled nitric oxide (FENO) and lung function
were not unexpected; epidemiologic data show that airway inflammation indicated by FENO does not
correlate strongly with either respiratory symptoms or lung function (Smith and Taylor, 2005). The
authors conclude that lung function decrements may be associated with the combustion-generated
component of ambient PM2 5, whereas airway inflammation may be related to some other constituent of
the ambient PM2 5 mixture.
Several studies outside of the U.S. examined eNO in relation to PM exposure among children.
Fischer et al. (2002) and Murata et al. (2007) found a significant association between increases in PM and
increases in the percent of eNO. Holguin et al. (2007) found no association between exposure to PM and
eNO. However, they did see significant associations between increases in eNO for the 95 asthmatic
subjects and measures of road density of roads 50- and 75-m from the home.
December 2008
6-129
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Exhaled Nitrogen Oxide - Adults
Three recent panel studies examined the effects of particle exposure on eNO measured in older
adults (Adamkiewicz et al., 2004) in Steubenville; (Jansen et al., 2005) in Seattle; Adar et al. 2007 in St.
Louis). Mean concentration data from these studies are characterized in Table 6-7. Breath samples were
collected weekly for 12 weeks from a group of 29 elderly adults in Steubenville (Adamkiewicz et al.,
2004). In single-pollutant models, each 10 (ig/m3 increase in 24-h ambient PM2 5 increased eNO by 0.82
ppb (95% CI: 0.19-1.45), a change of 15% compared to mean eNO (9.9 ppb). Effects were essentially
unchanged in multipollutant models that included ambient and/or indoor NO. The effect estimates for the
7 COPD subjects were significantly higher than for normal subjects (2.20 vs. 0.45 ppb, p = 0.03)
(Adamkiewicz et al., 2004).
In the Seattle panel of older adults (aged 60-86 years), 7 subjects were asthmatic and 9 had a
diagnosis of COPD (5 with asthma and 4 without) (Jansen et al., 2005). Exhaled NO was measured daily
for 12 days, along with personal, indoor, outdoor and central site PMi0, PM2 5 and BC. Significant
associations between 24-h average PM and eNO were found only for the asthmatic subjects: 10 (ig/m3
increases in outdoor levels (measured outside the subjects' homes) of PM2 5 or PMi0 were associated with
increases in eNO of 4.23 ppb (95% CI: 1.33-7.13), an increase of 22% above the group mean of 19.2 ppb,
and 5.87 ppb (95% CI: 2.87-8.88), an increase of 31%, respectively. BC measured indoors, outdoors or
personally was also associated with significant increases in eNO (of 3.97, 2.32, and 1.20 ppb,
respectively) (Jansen et al., 2005).
Adar et al. (2007b) conducted a panel study of 44 non-smoking seniors residing in St. Louis, MO
(age = 60 years). As part of the study, subjects were taken on group trips to a theater performance, Omni
movie, outdoor band concert, and a Mississippi River boat cruise. Subjects were driven to and from each
event aboard a diesel bus. Before and after each bus trip, eNO was measured on each subject. Two carts
containing continuous air pollution monitors were used to measure group-level micro-environmental
exposures to PM2 5, BC, and size-specific particle counts (0.3-2.5 |_im and 2.5-10 |_im) on the day of each
trip. Each 10 (ig/m3 increase in 24-h mean PM2 5 concentration was associated with a 36% increase in
eNO pre-trip (95% CI: 5-71). Each 10 (ig/m3 increase in micro-environmental PM2 5 concentration (i.e.
during the bus ride) was associated with a 27% increase in eNO post-trip (95% CI: 17-38).
These studies all demonstrated an association between increased levels of eNO and increases in
PM in the previous 4 to 24-h. Further, three studies demonstrated effects in elderly populations
(Adamkiewicz et al., 2004; Adar et al., 2007b; Jansen et al., 2005) while four others reported a similar
acute increase in eNO among children (Delfino et al., 2006; Koenig et al., 2003; Koenig et al., 2005;
2005a).
December 2008
6-130
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Outside of the U.S., one study examined eNO in a panel of 60 adult asthmatic subjects in London.
McCreanor et al. (2007) reported that 1 (ig/m3 increase in personal ezposure to EC was associated with
significant increases of approximately 1.75 to 2.25 % in eNO (results were presented graphically only) for
up to 22 h post-exposure.
Other Biomarkers of Pulmonary Inflammation
Other biomarkers of respiratory distress that have been examined in recent panel studies include
urinary leukotriene E4 (LTE4) in asthmatic children (Rabinovitch et al., 2006); and breath acidification in
adolescent athletes (Ferdinands et al., 2008). Mean concentration data from these studies are characterized
in Table 6-7.
In Rabinovitch et al. (2006), LTE4, an asthma-related biological mediator, was used to study the
response to short-term particle exposure. In the second winter of their 2-year study of asthmatic children
(described above in Section 6.3.1.1., under respiratory symptom and medication use outcomes), urine
samples were collected at approximately the same time of day from 57 subjects for 8 consecutive days.
Controlling for days with URI symptoms, each 10 (ig/m3 increase in morning maximum PM2 5 (measured
by TEOM), was associated with an increase in LTE4 levels by 5.1% (95% CI: 1.6-8.7). No significant
effects were observed on the same day or up to 3 days later based on 24-h averaged concentrations from
the TEOM monitor or from the FRM central site monitor.
The effects of vigorous outdoor exercise during peak smog season in Atlanta, GA on breath pH, a
biomarker of airway inflammation, in adolescent athletes (n = 16, mean age = 14.9 years) were examined
by Ferdinands et al. (2008). Median pre-exercise breath pH was 7.58 (range 4.39-8.09) and median
post-exercise breath pH was 7.68 (range 3.78-8.17). The authors observed no significant association
between ambient PM and post-exercise breath pH. However both pre- and post-exercise breath pH were
strikingly low in these athletes when compared to 14 relatively sedentary healthy adults and to published
values of breath pH in healthy subjects. The authors speculate that repetitive vigorous exercise may
induce airway acidification.
6.3.3.2. Human Clinical Studies
Studies of intrapulmonary instillation of particles in human subjects have provided evidence of
lung inflammation induced by exposure to PM. Lay et al. found that instillation of iron oxide (2.6 |im)
produced an increase in alveolar macrophages and neutrophils in BAL fluid collected 24-h
post-instillation. Ghio and Devlin (2001) evaluated the inflammatory response following instillation of
particles extracted from filters collected in the Utah Valley both prior to and after the closure of an area
steel mill. Subjects who underwent pulmonary instillation of particles (500 |ig) collected while the steel
December 2008
6-131
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
mill was operating (n = 16) had significantly higher levels of neutrophils 24-h post-instillation compared
with either saline instillation or with subjects (n = 8) who were instilled with the same mass of PM
collected during the mill's closure. This finding indicates that metals may be an important PM component
for this health outcome. A similar study by Schaumann et al. (2004) investigated the inflammatory
response of human subjects instilled with PM2.5 (100 |ig) collected from two different cities in Germany,
Hettstedt and Zerbst. Although instillation of PM from both cities were shown to induce airway
inflammation, instillation of PM from the more industrial area (Hettstedt) resulted in greater influxes of
BALF monocytes compared to PM collected from Zerbst. The authors postulated that the difference in
response between PM from the two cities may be due to the higher concentration of transition metals
observed in the samples collected from Hettstedt. In an inhalation study of exposure to fine CAPs
(23-311 |ig/m3) from Chapel Hill, NC, Ghio et al. (2000) observed an increase in airway and alveolar
neutrophils 18 hours after the 2-h exposure. Huang et al. (2003c) reported the increase in BAL neutrophils
demonstrated by Ghio et al. (2000) to be positively associated with the Fe, Se, and sulfate content of the
particles.
Samet et al. (2007) summarized the findings of Ghio et al. (2000) and presented preliminary data
from two studies that evaluated health effects of controlled 2-h exposures to ultrafine and thoracic coarse
PM. Ultrafine CAPs (47 |ig/m3) did not alter markers of pulmonary inflammation measured in BAL fluid
collected 18 hours after exposure. Pietropaoli et al. (2004) also observed a lack of airway inflammatory
response 21 hours after exposure to ultrafine carbon particles (10-50 |ig/m3). However, as described in
Samet et al. (2007), thoracic coarse CAPs (89 (ig/m3) were shown to significantly increase the percentage
of polymorphonuclear leukocytes (PMNs) in BAL fluid 20 hours after exposure. No thoracic coarse
PM-induced changes in macrophages, lymphocytes, monocytes, or eosinophils were observed. Alexis et
al. (2006) recently evaluated the effect of PM10-2.5 on markers of airway inflammation, specifically
focusing on the impact of biological components of PMi 0-2.5. Healthy men and women (n = 9) between the
ages of 18 and 35 inhaled nebulized saline (0.9%) as well as aerosolized PM10-2.5 collected from ambient
air. Subjects were exposed to PMi0-2.5 on two separate occasions, once using PMi0-2.5 that had been heated
to inactivate biological material and once using non-heated PMi0_2.5. Approximately 0.65 mg PMi0_2.5 was
deposited in the respiratory tract of subjects during the exposures. Markers of inflammation and immune
function were analyzed in induced sputum collected 2-3 hours after inhalation of saline or PMi0.2.5. Both
heated and non-heated PMi0_2.5 were observed to increase the neutrophil response compared with saline.
Exposure to non-heated PMi0_2.5 was found to increase levels of monocytes, eotaxin, macrophage TNF-a
mRNA, and was also associated with an upregulation of macrophage cell surface markers. No such
effects were observed following exposure to biologically inactive PMi0_2.5. These results suggest that
while thoracic coarse fraction PM-induction of neutrophil response is not dependent on biological
December 2008
6-132
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
components, heat sensitive components of coarse PM (e.g., endotoxin) may be responsible for
PM-induced alveolar macrophage activation.
Barregard et al. (2008) examined the effect of a short-term exposure (4 hours) to wood smoke
(240-280 |ig/m3) on markers of pulmonary inflammation in a group of healthy adults. Exposure to wood
smoke increased alveolar NO compared to filtered air (2.0 ppb versus 1.3 ppb) 3 hours after exposure.
Although these results provide some evidence of a PM-induced increase in pulmonary inflammation, the
physiological significance of the relatively small increase in alveolar NO is unclear. Larsson et al. (2007)
exposed 16 healthy adults to air pollution in a road tunnel for 2-h during the afternoon rush hour in
Stockholm, Sweden. The median PM2 5 and PMi0 concentrations during the road tunnel exposures were
64 |ig/m3 and 176 |ig/m3. respectively. Bronchial biopsies were obtained and broncoscopy and
bronchoalveolar lavage were performed 14 hours after the exposure. The results were compared with a
control exposure which consisted of exposure to urban air during normal activity. The authors reported
significant BALF increases in lymphocytes, total cell number, and alveolar macrophages following
exposure to road tunnel exposure versus control. No changes in adhesion molecules or blood coagulation
factors were observed. These results provide evidence of a significant association between exposure to
road tunnel air pollution and airway inflammation. However, unlike other controlled exposure studies, the
control exposure was not a true clean air control, but only a lower dose exposure group. In addition, it is
not possible to separate out the contributions of each air pollutant, including PM, on the observed
inflammatory response.
In a recent study evaluating the effect of DE exposure on markers of airway inflammation, Behndig
et al. (2006) exposed healthy adults (n = 15) for 2-h with intermittent exercise to filtered air or DE with a
PM10 concentration of 100 (.ig/m3. Eighteen hours after exposure to DE, the authors found significant
increases in neutrophil and mast cell numbers in bronchial tissue, as well as significant increases in
neutrophil numbers and IL-8 in bronchial lavage fluid compared with filtered air control. Similarly,
Stenfors et al. (2004) observed an increase in pulmonary inflammation (e.g., airways neutrophilia and an
increase in IL-8 in bronchoalveolar lavage fluid) among healthy adults 6 h following exposure to DE
(PM10 average concentration 108 |ig/m3). It is interesting to note, however, that no such inflammatory
effects were observed in a group of mild asthmatic subject in the same study. The diesel exhaust-induced
neutrophil response in the airways of healthy subjects observed in these two studies (Behndig et al., 2006;
Stenfors et al., 2004) is qualitatively consistent with the findings of Ghio et al. (2000) who exposed
healthy subjects to Chapel Hill fine CAPs. Another study reported no change in inflammatory markers in
nasal lavage fluid 4 and 96 h following intranasal instillation of DEP (300 (ig/nostril) in asthmatics and
healthy adults (Kongerud et al.). Pre-exposure of DE particles to ozone was not shown to have any effect
on the response. Although not a cross-over design, these findings suggest that exposure to DE particles
December 2008
6-133
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
without the gaseous component of DE may have little effect on inflammatory responses in human
subjects. In a group of healthy volunteers, Bosson et al. (2007) demonstrated that exposure to ozone (2 h
at 0.2 ppm) may enhance the airway inflammatory response of DE relative to clean air (1-h exposure to
300 |ig/m3). Exposure to DE was conducted h after exposure to ozone, and resulted in an increase in the
percentage of neutrophils in induced sputum collected 18 h after DE exposure. In a subsequent study
using a similar protocol at the same concentrations, DE was shown to increase the inflammatory effects of
ozone exposure, demonstrated as an increase in neutrophil and macrophage numbers in bronchial wash
(Bosson et al., 2008).
These new studies strengthen the evidence of PM-induced pulmonary inflammation, with the
majority of the evidence associated with fine and thoracic coarse fractions. Several studies suggest that
metal components of PM2.5 may be responsible for inflammatory responses (Ghio and Devlin, 2001;
Huang et al., 2003c; Schaumann et al., 2004). Others have reported inflammatory responses to DEP.
Differenct inflammatory responses were found with heat-sensitive and non-heat-sensitive fractions of
PMio.2.5 (Alexis et al., 2006).
6.3.3.3. Toxicological Studies
The 2004 PM AQCD discussed numerous studies investigating pulmonary inflammation in
response to CAPs, ROFA, DEPs, metals and acid aerosols. A wide variety of responses was reported
depending on the type of PM and route of administration. In general, exposure to fly ash and metal PM by
intratracheal instillation resulted in notable pulmonary inflammation. In contrast, inhalation of sulfates
and acid aerosols had minimal if any effect on pulmonary inflammation. More recent animal toxicological
studies using CAPs, DE and other relevant PM types are summarized below.
CAPs Studies
The 2004 PM AQCD found that fine CAPs exposure of rats and dogs at concentrations of
100-1000 ug/m3 for 1-6 h/day and 1-3 days generally resulted in minimal to mild inflammation in healthy
animals. Somewhat enhanced inflammation was observed in a model of chronic bronchitis. Since the last
review, numerous studies have investigated inflammatory responses to fine and ultrafine CAPs in both
healthy and compromised animal models.
In one study of healthy animals, Sprague Dawley rats were exposed to CAPs for 4 h/day on 3
consecutive days in Fresno, CA, in fall 200 and winter 2001, (PM2 5; mean mass concentration
190-847 |_ig/m3) (Smith et al., 2003). The particle concentrator used in these studies was capable of
enhancing the concentration of ultrafine as well as fine particles. Immediately after exposure on the third
day, BALF was collected and analyzed for total cells and neutrophils. Statistically significant increases
December 2008
6-134
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
were observed in numbers of neutrophils during the first week of the fall exposure period and in numbers
of total cells, neutrophils and macrophages during the first week of the winter exposure period. CAPs
concentrations were >800 (.ig/ni3 during both or those weeks.
Two studies were conducted using the CAPs in Boston. In a study by Godleski et al. (2002),
healthy Sprague Dawley rats were exposed for 5 h/day for 3 consecutive days to CAPs ranging in
concentration from 73.5-733.0 (ig/m3. BALF and lung tissue were collected for analysis 1 day later.
Neutrophilic inflammation was indicated by a statistically significant increase in percent neutrophils in
BALF. Microarray analysis of RNA from lung tissue and BAL cells demonstrated increased gene
expression of proinflammatory mediators, markers of vascular activation and enzymes involved in
organic chemical detoxification. This study overlapped in part with previously described studies by
Saldiva et al. (2002, discussed in the 2004 PM AQCD) and Batalha et al. (2002); see Section 6.2.4.3.). In
another study, healthy Sprague Dawley rats were exposed for 5 h to CAPs (mean mass concentration
1228 (ig/m3; 6/20-8/16/2002; (Rhoden et al., 2004). A statistically significant increase in BALF
neutrophils was observed 24-h following CAPs exposure. Histological analysis confirmed the influx of
inflammatory cells (Section 6.3.5.2.). Inflammation was accompanied by injury which is discussed in
Section 6.3.5.2.
Kodavanti et al. (2005) reported two sets of studies involving fine CAPs exposure during fall
months in RTP, NC. In the first study, SH rats were exposed to filtered air or CAPs (mean mass
concentration range 1,138-1,765 (ig/m3; less than 2.5 (.un in size) for 4 h and analyzed 1-3 h later. No
increase in BALF inflammatory cells or other measured parameter was observed. In the second study, SH
and WKY rats were exposed to filtered air or CAPs (mean mass concentration range 144-2,758 (ig/m3;
less than 2.5 (.im in size) for 4 h/day on 2 consecutive days and analyzed 1 day afterward. Differences in
baseline parameters were noted for the two rat strains since SH rats had greater numbers of BALF
neutrophils than WKY rats. Following the 2-day CAPs exposure, increased BALF neutrophils were
observed in the WKY rats but not in the SH rats compared with filtered air controls. Inflammation was
accompanied by an increase in a BALF marker of injury (see Section 6.3.5.2.).
Two CAPs studies of SH rats were conducted in the Netherlands. In the first study, SH rats were
exposed by nose-only inhalation to CAPs (ranging in concentration from 270-3660 |_ig/m3and in size from
0.15-2.5 |_im) from 3 different sites in the Netherlands (suburban, industrial and near-freeway) for 6 h
(Cassee et al., 2005). Increased numbers of neutrophils were observed in BALF 2 d post-exposure
compared to air controls. When CAPs exposure was used as a binary term, the relationship between CAPs
concentration and number of PMN in BALF was statistically significant. In contrast, Kooter et al. (2006)
reported no changes in markers of pulmonary inflammation measured 18 h after a 2-day exposure
(6 h/day) of SH rats to fine or fine+ultrafine CAPs from sites in the Netherlands (mean mass
December 2008
6-135
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
concentration range 399-3613 and 269-556 |ig/nr\ respectively; fine CAPs site in Bilthoven and
ultrafine+fine site in freeway tunnel in Hendrik Ido Ambacht).
Pulmonary inflammation was investigated in 2 studies using a rat model of pulmonary
hypertension (i.e., Sprague Dawley rats pre-treated with monocrotaline). In the first study, rats were
exposed to fine CAPs from an urban high traffic area in Taiwan (mean mass concentration of 371 (.ig/m')
(Lei et al., 2004b) for 6 h/day on 3 consecutive days and BALF was collected 2 days later. A statistically
significant increase in total cells and neutrophils was observed in BALF. Levels of TNFa and IL-6 in the
BALF were not altered by CAPs exposure. In the second study, rats were exposed to fine CAPs (mean
mass concentration 315.6 and 684.5 (.ig/m3 for 6 and 4.5 h, respectively; Chung-Li area, Taiwan) during a
dust storm event occurring 3/18-3/19/2002 (Lei et al., 2004b). Only one animal served as control during
the 6 h exposure (from 2100-300 on the first exposure day) and the data were combined with 3 control
animals from the 4.5 h exposure (from 300-730) on the second exposure day. A statistically significant
increase in total cells and neutrophils in BALF occurred in both CAPs-exposed groups. In addition,
increases in BALF IL-6 and markers of injury (see Section 6.3.5.2.) were observed as a function of CAPs
exposure.
In summary, pulmonary inflammation was noted in all 3 studies involving multi-day exposure of
healthy rats to CAPs from different locations. No pulmonary inflammation was seen in one study of SH
rats exposed to CAPs for 4 h and analyzed 1-3 h later. In studies involving multi-day exposure of SH rats,
two demonstrated pulmonary inflammation while one did not. In the rat monocrotaline model of
pulmonary hypertension, both single-day and multi-day exposures to CAPs resulted in mild pulmonary
inflammation.
On-Road Exposures
In a study by Elder et al. (2004a), old Fisher 344 rats (21 mo) were exposed to on-road highway
aerosols (particle concentration range 0.95-3.13 /10" particles/cm3; mass concentration estimated to be
37-106 (ig/m3; Interstate 90 between Rochester and Buffalo, NY) for 6 h on 1 or 3 consecutive days. No
increase in BALF inflammatory cells was observed 18 h post-exposure in any of the treatment groups.
Urban Air Studies
To evaluate inflammatory responses to ambient particles from vehicles, Wistar rats were exposed to
ambient urban air from a high traffic site (concentration range 22-225 (ig/m3 PMi0; Porto Alegre, Brazil)
or to the same air which was filtered to remove the PM (Pereira et al., 2007). Concentrations of gases
were not reported. Compared with filtered air controls, a significant increase in total number of BALF
December 2008
6-136
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
cells was observed 24 h following the 20 h continuous exposure but not following the 6 hours of exposure
to unfiltered urban air.
Diesel Exhaust Studies
The 2004 PM AQCD summarized findings of the 2002 EPA Diesel Document regarding the health
effects of DE. Short-term inhalation exposure to low levels of DE results in the accumulation of DPM in
lung tissue, pulmonary inflammation and alveolar macrophage aggregation and accumulation near the
terminal bronchioles. More recent studies are summarized below.
Pulmonary inflammatory responses were investigated in C57BL/6 mice exposed to diesel engine
emissions 7 h/d for 6 consecutive days (Harrod et al., 2003). Compared with controls, inflammatory cell
counts in BALF were increased in mice exposed to the higher concentration of DE (1000 (ig/m3 DEP) but
not in mice exposed to the lower concentration of DE (30 |_ig/m3 DEP). Concentrations of gases present in
the higher dose DE were reported to be 43 ppm NOx, 20 ppm CO and 364 ppb S02.
In a second study evaluating DE effects on BALF inflammatory cells, no increases in numbers of
neutrophils, lymphocytes or eosinophils were observed in BALB/c mice exposed by inhalation to 500 or
2000 (ig/m3 DEP for 4 h/d on 5 consecutive days (Stevens et al., 2008). Concentrations of gases reported
in this study at the higher concentration were 4.2 ppm CO, 9.2 ppm NO, 1.1 ppm N02, and 0.2 ppm S02
for the higher concentration of DE. Transcriptional microarray analysis demonstrated upregulation of
chemokine and inflammatory cytokine genes, as well as genes involved in growth and differentiation
pathways, in response to the higher concentration of DE. No gene expression results were reported for the
lower concentration of DE. Sensitization and challenge with ovalbumin significantly altered these
findings (see Section 6.3.6.2.). These results demonstrate that changes in gene expression can occur in the
absence of measurable pulmonary inflammation or injury markers (see Section 6.3.5.2.).
Li et al. (2007) exposed BALB/c and C57BL/6 mice to clean air or to low dose DE (DEP
100 (.ig/ni3) for 7 h/day and 5 days/week for 1, 4 and 8 weeks. Concentrations of gases in the DE were
reported to be 3.5 ppm CO, 2.2 ppm N02 and less than 0.01 ppm S02. Analysis of BALF and histology of
lung tissues was carried out at day 0 and after 1, 4 and 8 weeks of exposure. Total numbers of cells and
macrophages in BALF were significantly increased in C57BL/6 mice but not in BALB/c mice after 1
week exposure to DE compared with 0 day controls. Neutrophils and lymphocytes were increased after 1
week exposure to DE in both strains compared with 0 day controls. Differences in BALF cytokines were
also noted between the 2 strains after 1 week exposure to DE. No changes were observed by histological
analysis. Pulmonary function and oxidative responses were also evaluated (Sections 6.3.2.3. and 6.3.4.2.)
Long-term exposure responses are discussed in Sections 7.3.2.2, 7.3.3.2 and 7.3.4.1.
December 2008
6-137
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Healthy Fisher 344 rats and A/J mice were exposed to 30, 100, 300 and 1000 (ig/m3 DE PM by
whole body inhalation for 6 h/day, 7 days/week for either 1 week or 6 months in a study by Reed et al.
(2004). Concentrations of gases were reported to be from 2.0-45.3 ppm NO, 0.1-4.0 ppm N02,
1.5-29.8 ppm CO and 8-365 ppb for S02 in these exposures. One week of exposure resulted in no
measurable effects on pulmonary inflammation. Long-term exposure responses are discussed in Section
7.3.3.2.
In a study by Wong et al. (2003) and Witten et al. (2005), Fisher 344/NH rats were exposed
nose-only to filtered room air or to DE at concentrations of 35.3 |_ig/m3and 669.3 (ig/m3 DEP (particle size
range 7.2-294.3 nm) for 4 h/day and 5 days/week for 3 weeks. Gases associated with the high dose
exposure were reported to be 3.59 ppm NO, 3.69 ppm NOx, 0.1 ppm N02, 2.95 ppm CO, 518.96 ppm
carbon dioxide and 0.031 ppm total hydrocarbon. The focus of this study was on the possible role of
neurogenic inflammation in mediating responses to DE. Neurogenic inflammation is characterized by
both the influx of inflammatory cells and plasma extravasation into the lungs following the release of
neuropeptides from bronchopulmonary C-flbers. Pulmonary inflammation was evaluated by histological
analysis of lung tissue at the end of the 3 week exposure period. Following high, but not low,
dose-exposure to DE, a large number of alveolar macrophages was found in the lungs. Small black
particles, presumably DEP, were found in the cytoplasm of these alveolar macrophages. Perivascular
cuffing consisting of mononuclear cells was also observed in high dose-exposed animals. Influx of
neutrophils or eosinophils was not seen although mast cell number was increased in high-dose exposed
animals. Pulmonary plasma extravasation was measured by the "mTechnecium-albumin technique and
found to be dose-dependently increased in the bronchi and lung parenchyma. Alveolar edema was also
observed by histopathology in high dose-exposed animals. A significant decrease in Substance P content
in lung tissue was reported in DE-exposed rats. These responses initially suggested that DE resulted in
stimulation of C-fibers and activation of a local neuron reflex resulting in the repeated release of the
stored neuropeptide Substance P. Subsequent experiments were conducted using capsaicin pretreatment,
which inhibits neurogenic inflammation by activating C-fibers and causing the depletion of neuropeptide
stores. Pretreatment with capsaicin was found to reduce the influx of inflammatory cells but not plasma
extravasation in response to DE. Hence, DE is unlikely to act through bronchopulmonary C-fibers to
cause neurogenic inflammation in this model, although there may be a different role for
bronchopulmonary C-fibers in mediating the inflammatory cell influx.
Stimulation of bronchopulmonary C-fibers can result in activation of both local and CNS reflexes
through vagal parasympathetic pathways. McQueen et al. (2007) investigated the role of vagally-mediated
pathways in acute inflammatory responses to DEP. A statistically significant increase in BAL neutrophils
was observed 6 h after treatment of anesthetized Wistar rats with 500 |_ig DEP (SRM2975) by
December 2008
6-138
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
intratracheal instillation. This response was blocked by sectioning the vagus nerves or pretreatment with
atropine (McQueen et al., 2007). Similarly, atropine treatment blocked the increase in BAL neutrophils
seen 6 h after DEP exposure in conscious Wistar rats. These results provide evidence for the involvement
of a pulmonary vagal reflex in the inflammatory response to DEP.
In summary, several studies demonstrate that short-term inhalation exposure to DE
(100-1000 |_ig/m3 DEP) causes pulmonary inflammation in rodents. No attempt was made in these studies
to determine whether the responses were due to PM components or to gaseous components. However, PM
from DE was found to be capable of inducing an inflammatory response, as demonstrated by the one
intratracheal instillation study described above. Evidence was presented suggesting that DEP may act
through bronchopulmonary C-fibers to stimulate pulmonary inflammation.
Gasoline Emissions and Road Dust
Healthy male Swiss mice were exposed to gasoline exhaust (635 (ig/m3 PM and associated gases)
or filtered air for 15 min/day for 7, 14, and 21 days (Sureshkumar et al., 2005). BALF fluid was collected
for analysis 1-h after the last exposure. Histological analysis was also carried out at 7, 14, and 21 days.
The number of leukocytes in BALF fluid was increased after exposure to gasoline exhaust but this
increase did not achieve statistical significance. However, levels of the pro-inflammatory cytokines TNFa
and IL-6 were significantly increased in BALF following 14 and 21 days of exposure. Furthermore,
inflammatory cell infiltrate in the peribronchiolar and alveolar regions were observed by histology.
Evidence of lung injury was also found (see Section 6.3.5.2.). In this study, BALF analysis of
inflammatory cells was a less sensitive indicator of pulmonary inflammation than BALF analysis of
cytokines and histological analysis of lung tissue. Unfortunately results of this study cannot entirely be
attributed to the presence of PM in the gasoline exhaust since 0.11 mg/m3 SOx, 0.49 mg of NOx and 18.7
ppm of CO were also present during exposure.
Using ApoE"" mice on a high-fat diet, Campen et al. (2006) studied the impact of inhaled gasoline
emissions and road dust (6 h/day x 3 day) on pulmonary inflammation. Moreover, the investigators used a
high efficiency particle filter to compare the whole exhaust with an atmosphere containing only the
gaseous components. For gasoline emissions, the PM-containing atmosphere (PM mean concentration
61 |ig/m3: NOx mean concentration 18.8 ppm; CO mean concentration 80 ppm) failed to increase
numbers of inflammatory cells in BALF collected 18 h after the last exposure. However, a statistically
significant increase in total cells and macrophages was observed in response to resuspended road dust
(PM2 5) at 3500 (.ig/nr3. but not at 500 (ig/m3.
December 2008
6-139
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Fine and Ultrafine Carbon
In a study by Elder et al. (2004b), pulmonary inflammation was investigated in two compromised,
aged animal models (11-14 mo. SH and 23 mo. Fischer 344) exposed by inhalation to ultrafine carbon
black (count median diameter = 36 nm) at a relevant concentration (150 (.ig/ni3). No changes in BALF
cells were seen 24-h post-exposure in either model.
In a study by Gilmour et al. (2004c), adult Wistar rats were exposed for 7 h to fine and ultrafine
carbon black particles (mean mass concentration 1400 and 1660 (ig/m3 for fine and ultrafine CB,
respectively; mean number concentration 3.8* 103 and 5.2/104 particles/cm3, respectively; count median
aerodynamic diameter 114 nm and 268 nm, respectively). Both treatments resulted in increased BAL
neutrophils 16 h post-exposure, with the ultrafine particles having the greater response. Ultrafine particles
also increased total BALF leukocytes and macrophage inflammatory protein-2 mRNA in BALF cells.
Although these exposures may not be relevant to ambient exposures, this study demonstrated the greater
propensity of ultrafine carbon black particles to cause a proinflammatory response compared with fine
carbon black particles.
Iron-Soot
In a study by Last et al. (2004), BALB/c mice were exposed to 250 (ig/m3 laboratory-generated
iron-soot (size range 80-110 nm; about 200 (ig/m3 as soot) for 4 h/day and 3 days/week for 2 weeks.
BALF was collected 1-h after the last exposure and analyzed for total cells. No increase in total cell
number was observed following iron-soot exposure. Other findings of this study are described in Sections
6.3.2.3. and 6.3.5.2.
6.3.4. Oxidative Responses
The results of a small number of human clinical and toxicological studies presented in the 2004 PM
AQCD provided some initial evidence of an association between exposure to PM and pulmonary
oxidative stress. Recent human clinical studies have provided support to previous findings of an increase
in markers of pulmonary oxidative stress following exposure to diesel exhaust, and one new study has
observed a similar effect following controlled exposure to wood smoke. New findings from toxicological
studies also provide further evidence that oxidative species are involved in PM-mediated effects. No
epidemiological studies have evaluated the association between PM concentration and pulmonary
oxidative response.
December 2008
6-140
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
6.3.4.1. Human Clinical Studies
Two studies cited in the 2004 PM AQCD observed diesel-induced effects on airway oxidative
response following controlled human exposures (Blomberg et al., 1998; Nightingale et al., 2000). More
recently, Schaumann et al. (2004) demonstrated an increased oxidant radical generation of BAL cells
following instillation of urban particles compared with instillation of particles collected in a rural area.
The authors suggested that this difference was likely due to the greater concentration of transition metals
found in the urban particles. More recent studies have also evaluated the effect of exposure to WS and DE
on airway oxidative response in human subjects. Barregard et al. (2008) observed a significant increase in
malondialdehyde levels in breath condensate of healthy volunteers (n = 13) immediately following and 20
hours after a 4 hour exposure to WS (240-280 |ig/m3). Pourazar et al. (2005) exposed 15 adults (11 males
and 4 females) for 1-h to air or DE (PMi0 concentration 300 |ig/m3) in a controlled cross-over study.
Bronchoscopy with airway biopsy was performed 6 hours after exposure. The expression of NF-kB, AP-1
(c-jun and c-fos), p38, and JNK in bronchial epithelium was quantified using immunohistochemical
staining. DE was observed to significantly increase nuclear translocation of NF-kB, AP-1, phosphorylated
p38, and phosphorylated JNK; however, the findings of this study require confirmation with more
quantitative methods such as Western blot analysis. The observed activation of redox-sensitive
transcription factors by DE may result in the induction of proinflammatory cytokines. There is some
evidence to suggest that this bronchial response to DE is mediated through the epidermal growth factor
receptor signaling pathway (Pourazar et al., 2008). Behndig et al. (2006) evaluated the upregulation of
endogenous antioxidant defenses following exposure to DE (100 |ig/m3 PM10) in a group of 15 healthy
adults. Increases in urate and reduced glutathione were observed in alveolar lavage 18 hours after
exposure; however, no changes in urate or glutathione levels were observed in bronchial lavage. Taken
together, these studies suggest that short-term exposure to PM at near ambient levels may produce mild
oxidative stress in the lung. Limited data suggest that proximal and distal lung regions may be subject to
different degrees of oxidative stress during exposures to different pollutant particles.
6.3.4.2. Toxicological Studies
The 2004 PM AQCD reported one study which provided evidence that ROS were involved in
PM-mediated responses. This particular study used pre-treatment with the antioxidant DMTU to block the
neutrophilic response to ROFA. More recently, several studies evaluated the effects of PM exposure on
pulmonary oxidative stress. Oxidative stress can be directly determined by measuring ROS or oxidation
products of lipids and proteins. An indirect assay involves measurement of the enzyme heme oxygenase-1
(HO-1) or of the antioxidant enzymes superoxide dismutase or catalase, all of which can be induced by
December 2008
6-141
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
oxidative stress. Antioxidant interventions which inhibit or prevent responses are a further indirect
measure of oxidative stress playing a role in the pathway of interest.
Gurgueira et al. (2002) measured oxidative stress in Sprague Dawley rats immediately following a
5-h CAPs exposure (PM2 5; mean mass concentration range 99.6-957.5 |ig/m3: Boston, MA) and reported
increased in situ chemiluminescence (CL) in lungs of CAPs-exposed animals. CL evaluated after CAPs
exposure durations of 3 h was also increased but did not achieve statistical significance compared to the
filtered air group. When animals were allowed to recover for 24-h following the 5 h CAPs exposure, CL
levels returned to control values. Interestingly, a decrease in lung CL was observed in rats breathing
filtered air for 3 days compared with rats breathing room air for the same duration. Exposure to CAPs for
3 and 5 h also increased lung wet/dry ratios, indicating the presence of mild edema. To compare potential
particle-induced differences in in situ CL, rats were exposed to ROFA (1.7 mg/m3 for 30 min) or carbon
black (170 |ig/m3 for 5 h). Only the ROFA-treated animals exhibited increased CL in lung tissue.
Additionally, levels of antioxidant enzymes in the lung (MnSOD and catalase) were increased in
CAPs-exposed rats. A CAPs-associated increase in CL was also seen in the heart (see Section 6.2.9.3.) but
not the liver.
In a similar study, Rhoden et al. (2004) exposed Sprague Dawley rats for 5 h to CAPs from Boston
(mean mass concentration 1228 (.ig/ni3) or to filtered air. Significant increases in TBARS (a measure of
lipid peroxidation) and protein carbonyl content (a measure of protein oxidation) were observed 24-h
post-exposure to CAPs. Pretreatment with the thiol antioxidant N-acetylcysteine (NAC) (50 mg/kg ip) 1-h
prior to exposure prevented not only the lipid and protein oxidation observed in response to CAPs, but
also the increase in BALF neutrophils and and pulmonary edema in this model (see Sections 6.3.3.3. and
6.3.5.2.). Results of this study demonstrate the key role played by oxidative stress in these
CAPs-mediated effects.
Kooter et al. (2006) reported an increase in HO-1 in BALF and lung tissue measured 18 h after a
2-day exposure (6 h/day) of SH rats to fine or fine+ultrafine CAPs (mean mass concentration range
399-3613 and 269-556 |ig/m3. respectively; fine CAPs site in Bilthoven and ultrafine+fine site in freeway
tunnel in Hendrik Ido Ambacht). This occurred in the absence of any measurable pulmonary
inflammation (see Section 6.3.3.3.).
To evaluate oxidative stress responses to ambient particles from vehicles, Wistar rats were exposed
to ambient urban air from a high traffic site (concentration range 22-225 (ig/m3 PMi0; Porto Alegre,
Brazil) or to the same air which was filtered to remove the PM (Pereira et al., 2007). Several exposures
regimens were carried out: 6 and 20 h continuous exposures or to intermittent exposures of 5 h/day for 4
consecutive days. A significant increase in lipid peroxidation (measured as malondialdehyde) was seen in
December 2008
6-142
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
lung tissue immediately following the 20 h continuous exposure but not following the 6 h exposure or the
intermittent exposures.
Li et al. (2007) exposed BALB/c and C56BL/6 mice to clean air or to low dose DE (DEP
100 |_ig/m3) for 7 h/day and 5 days/week for 1, 4 and 8 weeks. Average gas concentrations were reported
to be 3.5 ppm CO, 2.2 ppm N02, and less than 0.01 ppm S02. HO-1 mRNA and protein were increased in
lung tissues of both mouse strains after 1 week of DE exposure. In addition, changes in airways
hyperresponsiveness and BAL cells and cytokines were observed (see Sections 6.3.2.3. and 6.3.3.3.).
Pretreatment with the thiol antioxidant NAC (320 mg/kg) IP on days 1-5 of DE exposure greatly
attenuated the increased airway hyperresponsiveness and increases in BAL inflammatory cells seen after
1 week of DE exposure. Long-term responses are discussed in Sections 7.3.2.2, 7.3.3.2 and 7.3.4.1.
A study by Whitekus et al. (2002) investigated the adjuvant effects of DEP in an allergic animal
model and is discussed in detail below (see Section 6.3.6.2.). Intervention with the thiol antioxidants
bucillamine and NAC inhibited the increase in allergen-specific IgE and IgGl as well as the increase in
protein carbonyl and lipid hydroperoxides in the lung following DE exposure.
6.3.5. Pulmonary Injury
The 2004 PM AQCD presented evidence from several toxicological studies of small PM-induced
increases in markers of pulmonary injury including thickening of alveolar walls and increases in
bronchoalveolar lavage fluid protein. These findings are consistent with the results of recent toxicological
studies demonstrating mild pulmonary injury accompanying inflammatory responses to CAPs. One recent
epidemiologic study has also observed a positive association between PM and urinary concentrations of
lung Clara cell protein.
6.3.5.1. Epidemiologic Studies
One epidemiologic study examined biomarkers of pulmonary injury. The mean concentration data
from this study are characterized in Table 6-7. Timonen et al. (2004) enrolled subjects with coronary heart
disease in Amsterdam (n = 37), Erfurt, Germany (n = 47) and Helsinki (n = 47) to study daily variation in
PM and urinary concentrations of lung Clara cell protein (CC16). No associations were seen between the
particle number concentration of the smallest particles (NC0 oi o 0 and CC16. Significant associations with
NC0 i-i and PM2 5 (which were strongly correlated with each other [r = 0.8]) were seen only for Helsinki
subjects: same day, lag 3 and 5-day mean NC01-1 increases of 1000/cm3 were associated with increases in
In (CC16/creatinine) of 15.5% (95% CI: 0.001-30.9), 17.4% (95% CI: 3.4-31.4), and 43.2% (95% CI:
17.4-69.0), respectively. Similar associations were seen for 10 (ig/m3 increases in PM2 5: lag 0 and 5-day
December 2008
6-143
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
mean PM2 5 were associated with increases in In (CC16/creatinine) of 23.3% (95% CI: 6.3-40.3) and
38.8% (95% CI: 15.8-61.8), respectively.
6.3.5.2. Toxicological Studies
The 2004 PM AQCD reported mild increases in BALF protein, a marker of pulmonary injury, in
several studies involving inhalation exposure to CAPs. In addition, histopathological analysis
demonstrated that the bronchoalveolar junction was the site of the greatest inflammation. Low level
exposure to DE was associated with Type 2 cell proliferation and thickening of alveolar walls near
alveolar macrophages according to the 2002 EPA Diesel Document. In addition, intratracheal instillation
of fly ash and metal-containing PM generally caused pulmonary injury as measured by increases in BALF
protein, LDH and albumin. Proliferation of bronchiolar epithelium was also noted. More recent studies of
BALF markers of pulmonary injury and histopathological analysis of lung tissue are summarized below.
BALF Markers of Pulmonary Injury and Increased Permeability
Kodavanti et al. (2005) exposed SH and WKY rats to filtered air or CAPs from RTP, NC (mean
mass concentration range 144-2,758 (ig/m3; less than 2.5 (.un in size) for 4 h/day on 2 consecutive days
and analyzed 1 day afterward. Differences in baseline parameters were noted for the two rat strains since
SH rats had greater levels of protein and lower levels of levels of LDH, NAG, ascorbate and uric acid in
the BALF than WKY rats. CAPs exposure resulted in increased levels of GGT in BALF (a marker of
epithelial injury) of SH rats but not WKY rats compared with filtered air controls. Injury was
accompanied by inflammation (see Section 6.3.3.3.).
In a study by Cassee et al. (2005), SH rats were exposed by nose-only inhalation to CAPs (ranging
in concentration from 270-3660 (ig/m3 and in size from 0.15-2.5 um) from 3 different sites in the
Netherlands (suburban, industrial and near-freeway) for 6 h. The pulmonary injury marker Clara Cell 16
protein (CC16) was increased in BALF following CAPs exposure. Inflammation was also observed (see
Section 6.3.3.3.).
Gurgueira et al. (2002) exposed Sprague Dawley rats to CAPs (PM2 5; mean mass concentration
range 99.6-957.5 |ig/m3: Boston, MA) and reported a small but statistically significant increase in lung
wet/dry ratios after 3 and 5 h of exposure, indicating the presence of mild edema. This response was
accompanied by increased oxidative stress as measured by in situ chemiluminescence (CL) (see Section
6.3.4.2). In a similar study, Rhoden et al. (2004) exposed Sprague Dawley rats for 5 h to CAPs from
Boston (mean mass concentration 1228 (.ig/ni3) or to filtered air. An increase in the wet/dry ratio (a
measure of edema) was observed 24-h following CAPs exposure which was diminished by pre-treatment
of the antioxidant N-acetylcysteine (see Section 6.3.4.2).
December 2008
6-144
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Pulmonary injury was investigated in 2 studies using a rat model of pulmonary hypertension
(Sprague Dawley rats pre-treated with monocrotaline; (Lei et al., 2004b). In the first study, rats were
exposed to CAPs from an urban high traffic area in Taiwan (mean mass concentration 371 (.ig/ni3) for 6
h/day on 3 consecutive days and BALF was collected 2 days later. A significant increase in BALF LDH
was observed in response to CAPs. In the second study, rats were exposed to ultrafine CAPs (mean mass
concentration 315.6 and 684.5 (ig/m3 for 6 and 4.5 h, respectively; Chung-Li area, Taiwan) during a dust
storm event occurring 3/18-3/19/2002 (Lei et al., 2004a). Only one animal served as control during the 6
h exposure (from 2100-300 on the first exposure day) and the data were combined with 3 control animals
from the 4.5 h exposure (from 300-730) on the second exposure day. Increases in BALF LDH and protein
were observed as a function of CAPs exposure. Pulmonary inflammation was observed in both of these
studies (see Section 6.3.3.3.).
In a study evaluating the effects of DE, no changes were observed in BALF protein and LDH in
BALB/c mice exposed by inhalation to concentrations of 50 and 2000 (ig/m3 DEP for 4 h/d on 5
consecutive days (Stevens et al., 2008). Concentrations of gases reported at the higher concentration of
DE were 4.2 ppm CO, 9.2 ppm NO, 1.1 ppm N02, and 0.2 ppm S02. Changes in gene expression were
measured in the higher exposure group. This study demonstrates that changes in gene expression can
occur in the absence of measurable markers of injury or pulmonary inflammation (see Section 6.3.3.3.)
In a study by Wong et al. (2003) and Witten et al. (2005), Fisher 344/NH rats were exposed
nose-only to filtered room air or to DE at concentrations of 35.3 |_ig/m3and 669.3 (ig/m3 DEP (particle size
range 7.2-294.3 nm) for 4 h/day and 5 days/week for 3 weeks. Gases associated with the high dose
exposure were reported to be 3.59 ppm NO, 3.69 ppm NOx, 0.1 ppm N02, 2.95 ppm CO, 518.96 ppm
carbon dioxide and 0.031 ppm total hydrocarbon. The focus of this study was on the possible role of
neurogenic inflammation in mediating responses to DE. Neurogenic inflammation is characterized by
both the influx of inflammatory cells and plasma extravasation into the lungs following the release of
neuropeptides such as Substance P from bronchopulmonary C-fibers. Pulmonary plasma extravasation
was measured by the "mTechnecium-albumin technique and found to be dose-dependently increased in
the bronchi and lung parenchyma. Alveolar edema was also observed by histological analysis of lung
tissue in high dose-exposed animals. A significant decrease in Substance P content in lung tissue was
reported in DE-exposed rats. These responses initially suggested that DE resulted in neurogenic
inflammation through the stimulation of bronchopulmonary C-fibers and activation of a local neuron
reflex resulting in the repeated release of the stored neuropeptide Substance P. Subsequent experiments
were conducted using capsaicin pretreatment, which inhibits neurogenic inflammation by activating
C-fibers and causing the depletion of neuropeptide stores. Pretreatment with capsaicin did not reduce
plasma extravasation following DE exposure. Hence, DE is unlikely to act through bronchopulmonary
December 2008
6-145
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
C-fibers to cause neurogenic inflammation in this model. Inflammatory responses measured in this study
are discussed in Section 6.3.3.3.
Healthy male Swiss mice were exposed to gasoline exhaust (635 |_ig/m3 PM and associated gases)
or filtered air for 15 min/day for 7, 14, and 21 days (Sureshkumar et al., 2005). BALF fluid was collected
for analysis 1-h after the last exposure. Statistically significant increases in BALF markers of lung injury,
alkaline phosphatase, gamma-glutamyl transferase and LDH, were observed at all time points studied.
Alveolar edema was noted following 14 and 21 d of exposure. Other findings of this study including
inflammation and histopathology are discussed in Section 6.3.3.3. Results of this study cannot entirely be
attributed to the presence of PM in the gasoline exhaust since 0.11 mg/m3 SOx, 0.49 mg of NOx and 18.7
ppm of CO were also present during exposure.
Histopathology
Histopathological changes were demonstrated in Sprague Dawley rats exposed for 5 h to CAPs
from Boston (mean mass concentration 1228 (ig/m3; (Rhoden et al., 2004)) compared with air-exposed
controls. Slight bronchiolar inflammation and thickened vessels at the bronchiole were observed 24-h
post-exposure. These findings are consistent with the influx of polymorphonuclear leukocytes observed in
BALF 24-h post-exposure to CAPs (see Section 6.3.3.3.).
An interesting study demonstrating histopathological responses to PM in neonatal rats was reported
by Pinkerton et al. (Pinkerton et al., 2004). Rat pups (10 day old) were exposed to soot and iron particles
(mean mass concentration of 243 (ig/m3; iron concentration 96 (ig/m3; size range 10-50 nm) for 6 h/day
on 3 consecutive days. Cell proliferation in different lung regions was evaluated following
bromodeoxyuridine injection 2-h prior to necropsy. The rate of cell proliferation in the proximal alveolar
region (immediately beyond the terminal bronchioles) was significantly reduced in iron-soot exposed
animals compared to controls. This was a region-specific response since the rate of cell proliferation was
not altered in the terminal bronchioles or the general lung parenchyma. However, alveolar septation and
growth was not found to be altered by iron-soot exposure. The authors suggest the possibility of greater
susceptibility to air pollution during the critical postnatal lung development period which occurs in
animals and humans and that neonatal exposure to PM may contribute to impaired lung growth seen in
children.
In another study investigating the effects of iron-soot, healthy adult BALB/c mice were exposed to
250 (ig/m3 laboratory-generated iron-soot (size range 80-110 nm; about 200 (ig/m3 as soot) for 4 h/day
and 3 days/week for 2 weeks (Last et al., 2004). Analysis of airway collagen content was conducted by
histology and by biochemical analysis of microdissected airways. No increases in airway collagen content
were found by either method in mice exposed to iron-soot for 2 weeks. Furthermore, no goblet cells were
December 2008
6-146
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
observed in airways of air or iron-soot exposed animals. Other findings of this study are described in
Sections 6.3.23.3. and 6.3.3.3.
Another study demonstrated histopathological responses to gasoline exhaust in healthy male Swiss
mice exposed to gasoline exhaust (635 (ig/m3 PM and associated gases) or filtered air for 15 min/day for
7, 14, and 21 days (Sureshkumar et al., 2005). Histological observations showed inflammatory cell
infiltrate in the peribronchiolar and alveolar region, alveolar edema and thickened alveolar septa at 14 and
21 days postexposure. Levels of pro-inflammatory cytokines and marker enzymes of lung damage were
also increased in BALF. The numbers of inflammatory cells in BALF was increased but not significantly,
demonstrating that BALF analysis of inflammatory cells was a less sensitive indicator of pulmonary
inflammation in this study than histophathological analysis. Results of this study cannot entirely be
attributed to the presence of PM in the gasoline exhaust since 0.11 mg/m3 SOx, 0.49 mg of NOx and 18.7
ppm of CO were also present during exposure.
In a study by Wong et al. (2003) and Witten et al. (Witten et al.), Fisher 344/NH rats were exposed
nose-only to filtered room air or to DE at concentrations of 35.3 (.ig/m'and 669.3 (ig/m3 DEP (particle size
range 7.2-294.3 nm) for 4 h/day and 5 days/week for 3 weeks. Gases associated with the high dose
exposure were reported to be 3.59 ppm nitric oxide, 3.69 ppm NOx, 0.1 ppm N02, 2.95 ppm CO, 518.96
ppm C02 and 0.031 ppm total hydrocarbon. Pulmonary inflammation was evaluated by histopathological
analysis of lung tissue. Following high, but not low, dose-exposure to DE, a large number of alveolar
macrophages was found in the lungs. Small black particles, presumably DEP, were found in the
cytoplasm of these alveolar macrophages. Perivascular cuffing consisting of mononuclear cells was also
observed in high dose-exposed animals. Influx of neutrophils or eosinophils was not seen although mast
cell number was increased.
Relative Toxicity of PM Size Fractions
Ambient PM Studies
A recently undertaken multinational project entitled "Chemical and biological characterization of
ambient thoracic coarse (PM10-2.5), fine (PM2.5-0.2), and ultrafine particles (PM02) for human health risk
assessment in Europe" (PAMCHAR) takes a systematic approach to expanding the present knowledge
about the physiochemical and toxicological effects of these three PM size fractions. Six European cities
were selected that represented contrasting ambient PM profiles: Helsinki, Duisburg, Prague, Amsterdam,
Barcelona, and Athens. For PM collected at all sites, PM10-2.5 induced the greatest pulmonary effects in
C57B1/6J mice intratracheally instilled with 1, 3, or 10 mg/kg of particles (Happo et al., 2007; 2006).
Dose-response relationships in BALF parameters measured 24-h post-instillation, including BALF cell
number and protein, were observed for all sites following PM10_2.5 instillation and neutrophils were the
December 2008
6-147
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
predominant cell type (Happo et al., 2007). Prague PM10_2.5 exposure resulted in decreased macrophages
in BALF at 12 h and Amsterdam, Barcelona, and Athens PM10.2.5 induced lymphoplasmacytic cells in
BALF (Happo et al., 2007). No inflammatory responses were observed for ultrafine PM measured 12-h
after exposure. Protein was elevated for PM10-2.5 at all locations with the 10 mg/kg dose; Athens ultrafine
PM induced protein release only at the two lowest doses 12 h post-exposure. For TNF-a and IL-6, the
greatest response was observed with PM10-2.5 4 h following exposure (Happo et al., 2007). Ultrafine
Duisburg PM exposure resulted in elevated TNF-a for the 1 and 3 mg/kg doses. Only the Helsinki sample
appeared to induce the same level of IL-6 release for PMi0_2.5 and PM0 2 at 10 mg/kg, albeit the collection
times differed. In vitro TNF-a and IL-6 responses did not always reflect in vivo effects (Table 6-8), as the
Duisburg PMi0.2.5 sample was the most potent in vivo compared to the other sites and elicited much lower
cytokine release compared to other cities (except Helsinki) in vitro (Happo et al., 2007; Jalava et al.,
2006).
Helsinki PM was collected in the spring and generally had the lowest in vivo and in vitro activity
for PM10-2.5 compared to the other cities (Happo et al., 2007; Jalava et al., 2007). Spring-time samples
were collected because episodes of resuspended road dust occur frequently during this season (Pennanen
et al., 2007). There was a high correlation between EC content in PM2 5 and PM10-2.5, indicating that traffic
impacted both size fractions (Sillanpaa et al., 2005). Duisburg PM collected in autumn had the greatest
amounts of Mn and Zn compared to PM samples from other locations (Pennanen et al., 2007). Metals
industries in Duisburg are likely contributors to the observed PM metals concentrations. For the Prague
winter PM samples, the As content was the higher than any other location (Pennanen et al., 2007). Prague
also had the highest PAH levels in all three size fractions, possibly attributable to stable atmosphere
conditions and incomplete combustion of coal and biomass in residential heating (Pennanen et al., 2007).
High levels of ammonium and nitrate in PM samples from Amsterdam suggest traffic as a large source of
air pollution (Pennanen et al., 2007). Approximately one-third of PMi0-2.5 mass from Amsterdam was
comprised of sea salt (Sillanpaa et al., 2006), double that of any other city. In Barcelona and Athens, high
calcium or Ca2+ contents in spring and summer PM2 5 and PM10-2.5 are indicative of resuspended
soil-derived particles (Pennanen et al., 2007).
December 2008
6-148
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Table 6-8.
PAMCHAR PM10-2.5 inflammation results with ambient PM.
City and Season
In vivo3 (mg/kg)
In vitrob (|jg/ml)
BALF protein BALF TNF-a
BALF IL-6
BALF KC
BALFPMN BALF AM
TNF-a
IL-6
MIP-2
Helsinki spring
+10 +10
+10
+3,10
+10
+150,300
+150,300
+150,300
Duisburg autumn
+10 +10
+10
+10
+10
+150,300
+150,300
+300
Prague winter
+10 +3,10
+10
+3,10
+10 +10
+150,300
+150,300
+150,300
Amsterdam winter
+10 +10
+10
+10
+10
+150
+150,300
+150,300
Barcelona spring
+10 +10
+3,10
+10
+10
+150,300
+150,300
+150,300
Athens summer
+10 +3,10
+3,10
+3,10
+10
+150,300
+150,300
+150,300
aSource: Happo et al. (2007); 2 cells used for in vitro study were RAW264.7
bSource: Jalava et al. (2007); + indicates increased response and numbers that follow indicate at which dose the response was observed
Another study employed Duisburg PM. In contrast to the PAMCHAR study where animals were
administered PM suspended in pathogen-free water (Happo et al., 2007), animals received PM via
intratracheal instillation suspended in saline at a dose of 320 (.ig (Schins and Knaapen, 2007). In female
Wistar rats, neutrophils in BALF were significantly elevated for PMi0.2.5 from Duisburg and Borken
(Table 6-9), albeit the percent of neutrophils with the PMi0.2.5 from Borken was nearly double that of
Duisburg (Schins et al., 2004). The responses with PM2 5 were much smaller. When these PMi0_2.5 particles
were introduced into whole blood to determine overall inflammogenic capacity, IL-8 and TNF-a were
released in greater quantities than in response to fine PM. Furthermore, PM10.2.5 from Borken induced
higher cytokine responses than Duisburg PM10.2 5.
An in vivo study conducted with SH rats that employed PM10_2.5 and PM2 5 from six different
European locations with varying traffic densities (3 or 10 mg/kg via intratracheal instillation; ultrafine PM
was not collected) reported that PM10_2.5 generally induced greater responses than PM2 5 (Gerlofs-Nijland
et al., 2007). PMi0.2 5 from a location with high traffic influence in Munich, Germany demonstrated the
greatest response in LDH activity, BALF protein, total cells, neutrophils, and lymphocytes 24-h
post-instillation (Gerlofs-Nijland et al., 2007). PMi0.2 5 collected from a low traffic site in Munich induced
the greatest cytokine response for TNF-a and MIP-2. Although some correlations were observed between
PMio_2 5 components (Ba and Cu) and BALF parameters, they were largely driven by one location
(Gerlofs-Nijland et al., 2007).
In an in vivo study that employed ambient PM collected in fall 1996 from Research Triangle Park
(RTP), NC, neutrophilic influx was observed in BALF of female CD1 mice 18-h post-instillation (10, 50
or 100 |ig) of coarse PM (3.5-20 |im), although a dose-response relationship was not evident (Dick et al.,
2003a). Only the two highest doses of PM for the smaller size fractions induced elevated neutrophils.
Significant responses in albumin and TNF-a were only observed for the fine PM (1.7-3.5 (j,m) exposure
group. Total protein, LDH, NAG, and responses were absent for all PM size fractions. Levels of IL-6 were
December 2008
6-149
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
elevated in mice exposed to 100 |ig for coarse, fine, and fine/ultrafine (< 1.7 |im) PM. When
dimethylthiourea (DMTU) was administered intravenously prior to exposure, the neutrophil response was
attenuated in all groups to levels below control.
Another study compared thoracic coarse, fine, and ultrafine PM collected in Seattle, WA, Salt Lake
City, UT, South Bronx, NY, and Sterling Forest, NY (Gilmour et al., 2007). In female BALB/c mice, the
100 |ig dose of PM10.2 5 (approximately 5 mg/kg) from Salt Lake City induced a significant increase in
protein in BALF and the level released was almost as high as that observed after LPS exposure (Gilmour
et al., 2007). PMi0.2.5 from the South Bronx resulted in dose-related increases in neutrophil number and
MIP-2 levels in BALF (Gilmour et al., 2007). The greatest amount of LPS was observed in the Salt Lake
City and Seattle PMio_2 5 samples. There was a less discernable patter of response with fine and ultrafine
PM.
Table 6-9. Other ambient PM - in vivo PM10-2.5 studies - BALF results, 18-24 h post-IT
Location
Endotoxin
D0Se CeN Cytokines lnjury
(mg/kg) Differentials " Biomarkers
Reference
Germany, Borken; rural Feb-May
2000
-6.6 EU/mg 0.58-0.91
f % PMN
t TNF-a
Schins et al. (2004)
Germany, Duisburg; heavy industry
Feb-May 2000
-5.0 EU/mg 0.58-0.91
t % PMN
t MIP-2
Schins et al. (2004)
USA, Seattle, WA
Feb-March 2004
-6.0 EU/mg 1.25,5.0
Gilmour, et al. (2007)
USA, Salt Lake City, UT
Apr-May 2004
-6.3 EU/mg 1.25,5.0
f protein
Gilmour, et al. (2007)
USA, South Bronx, NY
Dec 2003-Jan 2004
-2.8 EU/mg 1.25,5.0
t PMN
t MIP-2
Gilmour, et al. (2007)
USA, Sterling Forest, NY
Dec 2003-Jan 2004
-2.9 EU/mg 1.25,5.0
Gilmour, et al. (2007)
USA, RTP, NC
Oct-Nov 1996
-0.96 EU/mg 0.5, 2.5, 5.0 ft PMN
t IL-6
Dick, (2003a)
Germany, Munich Ost Bahnof; high
traffic A
Aug 2002
-2.9 EU/mg 3,10
f* total cells
t AM
fPMN
f* Lymph
ft MIP-2
ft TNF-a
ft* LDH
f* protein
Gerlofs-Nijland, et al.
(2007)
Netherlands, Hendrik-ldo-Ambacht;
high traffic
Sept 2002
3,10
-6.5 EU/mg
f total cells
t*AM
t PMN
f Lymph
t MIP-2
ft TNF-a
ft LDH
f protein
Gerlofs-Nijland, et al.
(2007)
Italy, Rome; high traffic
Apr 2002
-1.5 EU/mg 3,10
total cells
t AM
t PMN
f Lymph
ft MIP-2
ft TNF-a
ft LDH
Gerlofs-Nijland, et al.
(2007)
December 2008
6-150
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Location
Endotoxin
Dose
(mg/kg)
Cell
Differentials
Cytokines
Injury
Biomarkers
Reference
Netherlands, Dordrecht; moderate
traffic
Apr 2002
-0.6 EU/mg 3,10
ft total cells
t AM
ft PMN
f Lymph
ft LDH
f protein
Gerlofs-Nijland, et al.
(2007)
Germany, Munich Grosshadern
Hospital; low traffic
Jun—Jul 2002
-2.9 EU/mg 3,10
f total cells
ft AM
ft PMN
ft Lymph
ft* MIP-2
ft* TNF-a
ft* LDH
f protein
Gerlofs-Nijland, et al.
(2007)
Sweden, Lycksele; low traffic
Feb-March 2002
-0.9 EU/mg 3,10
ft total cells
t AM
ft PMN
f Lymph
ft LDH
f protein
Gerlofs-Nijland, et al.
(2007)
For Gerlofs-Nijland study, composition data were averaged across seasons, f significant only at highest dose, tt Significant at lowest and highest dose. * Greatest potency for that
endpoint and study. Gilmour et al. exposure was via aspiration.
Coal Fly Ash
Coal fly ash of differing size fractions and composition was injected into female CD1 mice
oropharynx (25 or 100 |ig) and then aspirated to assess lung inflammation and injury 18 h following
exposure (Gilmour et al., 2004a). Montana (low-sulfur subbituminous; 0.83% sulfur, 11.72% ash content)
or Western Kentucky (high-sulfur bituminous; 3.11% sulfur, 8.07% ash content) coal was combusted
using a laboratory-scale down-fired furnace. Interestingly, no significant PMi0.2.5 effects for either coal fly
ash were observed for BAL neutrophils, TNF-a, MIP-2, albumin, total protein, LDH activity, or NAG
activity 18 h post-IT. However, the ultraflne fraction (PM0 2) of combusted Montana coal induced greater
numbers of neutrophils than PMi0.2.5 or PM2 5 at both doses. TNF-a was only elevated in animals exposed
to 100 |ig of the Montana ultraflne PM; MIP-2 was also increased at both doses. The PM2 5 Western
Kentucky coal fly ash caused increased BAL neutrophils, MIP-2, albumin, and protein (Gilmour et al.,
2004a).
In a similar study employing Montana subbituminous coal fly ash particles >2.5 jjxn, C57B1/6J
mice were intratracheally instilled with PM alone or PM+LPS and BALF was obtained 18 h
post-exposure (Finnerty et al., 2007). TNF-a and IL-6 in lung homogenates were only elevated in the
animals exposed to PM+100 |ig LPS, although it appeared that there was a greater-than additive effect.
Total cells and cell differentials were not measured (Finnerty et al., 2007).
Summary
Biomarkers of injury and inflammation were measured in in vivo and in vitro studies comparing
the toxicity of different size fractions of ambient PM from various locations. Responses were measured in
BALF of rodents following intratracheal instillation or aspiration of PM. In general, the PMi0.2.5 size
December 2008
6-151
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
fraction was more potent than fine or ultrafine PM and endotoxin levels did not appear responsible. In one
study, rural PM10_2.5 from Germany induced a greater inflammatory and cytokine response than PM10_2 5
from an industrial location. In contrast, PM10.2.5 from Sterling Forest, NY did not lead to any change in
BALF inflammation or injury markers. A study that employed coal fly ash indicated that the ultrafine PM
fraction was the most inflammogenic. All of these studies were conducted using high doses of PM
(0.58-10 mg/kg) and it is unclear if similar effects would be observed at lower doses.
6.3.6. Allergic Responses
A large number of toxicological and human clinical studies cited in the 2004 PM AQCD reported
an exacerbation of existing allergic airway disease following exposure to laboratory-generated and
ambient particles. In addition, numerous studies have demonstrated that PM can alter the immune
response to challenge with specific antigens and suggest that PM may act as an adjuvant to promote
allergic sensitization. Recent toxicological studies have provided evidence of enhanced allergic response
and allergic sensitization following exposure to CAPs and diesel that is consistent with the findings
presented in the 2004 PM AQCD. Thus PM can enhance allergic responses by facilitating delivery of
allergenic material and promoting subsequent immune reactivity. The initiation or exacerbation of allergic
responses has important implications for allergic asthma, the most common form of asthma.
6.3.6.1. Human Clinical Studies
Exacerbation of Allergic Responses
Exposure to DE particles was shown to increase the allergic response among atopic individuals in
several human clinical studies cited in the 2004 PM AQCD. Nordenhall et al. (2001) found that exposure
to DE significantly decreased the concentration of methacholine required to induce a 20% decrease in
FEVi in a group of atopic asthmatics 24-h post-exposure. In addition, Diaz-Sanchez et al. (1997)
demonstrated an increase in allergen-specific IgE following exposure via intranasal spray to ragweed plus
DE particles relative to ragweed allergen alone. Decreases in IFN-y and IL-2, as well as increases in IL-4,
IL-5, IL-6, IL-10, and IL-13 were also observed when ragweed allergen was administered with exhaust
DE particles. One new study (Bastain et al., 2003) also observed an increase in IL-4 and allergen specific
IgE, as well as a decrease in IFN-y following intranasal administration of ragweed allergen with DE
particles in atopic adults. The protocol was repeated in this study for all subjects, and the enhancement of
allergic response by co-exposure to DE particles was observed to be highly reproducible within
individuals.
December 2008
6-152
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Allergic Sensitization
One controlled human exposure study has demonstrated that de novo sensitization to a neoantigen
can be induced by exposure to DE particles. In this study, Diaz-Sanchez et al. (1999) dosed 25 atopic
adults intranasally with 1 mg keyhole limpet hemocyanin (KLH), followed by 2 biweekly challenges with
100 |ig KLH. In 15 of the 25 subjects, DE particles were administered intranasally 24-h prior to each
KLH exposure, while in the other 10 subjects, no diesel particles were administered. No KLH-specific
IgE was observed in the nasal lavage fluid of any of the subjects exposed to KLH without exposure to
diesel particles. However, KLH-specific IgE was present in the nasal lavage fluid of 9 out of 15 subjects
28 to 32 days after the initial KLH immunization when exposures were preceded by administration of DE
particles.
6.3.6.2. Toxicological Studies
Exacerbation of Allergic Responses
Increased use of actual ambient air particle mixes in toxicological studies since the 2004 CD has
greatly expanded evidence relevant to assessing these and other immunotoxic effects. A number of studies
have also included ambient level doses, although many still include relatively high doses of questionable
relevance compared to the doses inhaled by humans. Recent dosimetric models reveal that a small
fraction of epithelial cells located at the carinal ridges of airway bifurcations can receive massive doses
that may be even a few hundred times higher than the average dose for the whole airway (see Chapter 4).
These areas, coincidentally, are locations of bronchus associated lymphoid tissues (BALT) which are sites
at which interaction of T and B lymphocytes with antigen presenting cells (APC) occurs. Hence the
deposited particles are in near-ideal proximity to immunologically active tissues. Doses used for assessing
PM immunotoxicity should be viewed with this perspective.
Existing allergic sensitization confers susceptibility to the effects of PM in rodent models. For
example, studies in allergic rats (Harkema et al., 2004; Morishita et al., 2004) suggest that allergic
sensitization enhances the retention of PM in the airways. OVA intranasally sensitized and challenged BN
rats were exposed to CAPs PM2 5 for 4 or 5 consecutive 10-h days. Exposures were conducted in Detroit
during July or September (4 or 5 d, time weighted average mass concentration of 676 ± 288 or
313 ± 119 (ig/m3, respectively). Recovery of anthropogenic trace elements (La, V, Mn, S) from lung tissue
24-h post-exposure was greater for CAPs exposed sensitized/challenged rats than for air exposed or
non-allergic CAPs exposed controls. Interestingly, temporal increases in these elements were associated
with eosinophil influx and BALF protein content, as well as significantly increased airway
mucosubstances, despite lower average mass concentration (September, 313 ± 119). During September,
December 2008
6-153
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
the average number concentration of ultrafine particles was nearly double that in July (10,879 ± 5126 vs.
5753 ± 2566 particles/cm3), and the authors speculated that this high concentration of ultrafine particles
facilitated particle penetration into the alveolar region of the lungs. Experiments were conducted using
intratracheal instillation of fractionated soluble and insoluble ambient PM2.5 collected during the same
time period during September (Harkema et al., 2004). A mild pulmonary neutrophilic inflammation was
observed in healthy BN rats instilled with the insoluble fraction of PM2 5, but instillation of total, soluble,
or insoluble PM2 5 in allergic rats did not result in differential effects.
Research has also been conducted to determine the effect of proximity to the roadway on
exacerbation of existing allergic disease. OVA-allergic BALB/c mice were exposed to CAPs (fine, F, <
2.5 or ultrafine, UF, < 0.15, avg. total concentration 400 (.ig/ni3) for five 4-h days a week over two weeks
at 50 or 150 m downwind of a heavily trafficked road (Kleinman et al., 2005). After two OVA inhalation
challenges, significantly higher concentrations of IL-5, OVA-specific IgE and IgGl, and eosinophils were
detected in serum and lavage samples from mice exposed to CAPs (UF or F) than in samples from
air-exposed mice. Mice exposed to CAPs closer to the roadway (50 m) had higher levels of IL-5, IgGl,
and eosinophils than mice exposed to CAPs 150 m downwind. The UF CAPs appeared to be more potent
in mediating these effects, and the authors suggest that the enhanced responses to exposure closer to the
roadway may reflect a greater proportion of UF particles in this vicinity, as the concentrations of
sub-25-nm particles decrease rapidly with distance from the roadway. Animal-to-animal variability among
the biomarkers tested made it necessary to combine values from two exposures spanning two years for
statistical power (determined prior to the start of the experiment). A subsequent publication (Kleinman et
al., 2007) included a third exposure regimen as well as compositional analysis. Fine CAPs mass
concentration was intentionally adjusted to an average concentration of approximately 400 |ig/m3. ranging
from 163 |ig/m3 to 500 (ig/m3. UF ranged from 146 to 430 |ig/nr\ Analysis of results from the three
exposures indicated that OVA-sensitized mice exposed 50 m downwind of the roadway exhibited
increased levels of IL-5 and IgGi compared to mice exposed 150 m downwind or exposed to air. No
markers of allergy-related responses were observed in the 150 m exposure groups, and very little
difference was seen between fine and ultrafine CAPs responses, perhaps because fine material contained
20-32% ultrafine components. The strongest associations between component concentrations and
biological markers of allergy (IL-5 and IgGl) were to EC and OC. This study suggests that proximity to a
source may be an important factor in allergic responses.
Existing allergic sensitization also modulates airway responses following exposure to PM. AHR, as
measured by methacholine-induced airway resistance, was observed in ovalbumin (OVA) sensitized
C57BL/6J mice after aerosol challenge with OVA and a single 5 hour nose-only exposure to a
concentration of 870 (ig/m3 aerosolized filter-collected DEP (PM2 5) (Farraj et al., 2006a). Intranasal
December 2008
6-154
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
pretreatment with an antibody against the pan neurotrophin receptor p75 attenuated the DEP-induced
increase in airflow obstruction, indicating a role for neurotrophins. Neurotrophins are expressed by
various structural, nerve and immune cells within the respiratory tract and are linked to the etiology of
asthma in both humans and animal models. DEP alone in unsensitized mice caused a significant increase
in lung macrophages; this response was also inhibited by anti-p75, which may suggest mediation of
macrophage influx by neurotrophin or alternatively may reflect anti-p75 dependent depletion of
macrophages due to expression of the p75 receptor. Aside from increased macrophages, the single
exposure to DEP had little effect on other markers of airway inflammation. In a similar subsequent study,
these authors demonstrate neurotrophin-mediated DEP-induced airflow obstruction in OVA sensitized and
challenged BALB/c mice (Farraj et al., 2006b), in this case using a higher 2000 |_ig/m3 single 5 h exposure
to aerosolized filter-collected PM2 5. Differences between whole body plethysmography and tracheal
ventilation measurements indicated that airflow obstruction may have originated in the nasal passages.
Again, very few indices of inflammation were increased; however, similar neurotrophin-dependent
increases in lung macrophages were observed after DEP exposure alone, and IL-4 protein levels were
increased 5-fold in the BALF of sensitized, challenged, DEP-exposed mice. This neurotrophin-dependent
IL-4 response was not evident in the first study, and may be related to the higher dose used in the second
study or the characteristic allergic or Th2 bias of the BALB/c strain. Airflow obstruction in the absence of
airway inflammation in OVA-sensitized animals seen in both studies by Farraj et al. (2006a, b) may
reflect DEP-induced acute enhancement of neurogenic as opposed to immunologic inflammation. In these
and other studies, particular effects such as airway obstruction are only evident when allergic sensitization
precedes DEP exposure.
Exposure to relatively low doses of DE has been shown to exacerbate asthmatic responses in OVA
sensitized and challenged BALB/c mice (Matsumoto et al., 2006). Mice were intraperitoneally sensitized
and intranasally challenged one day prior to chamber exposure to DE (DEP 100 (ig/m3; CO, 3.5 ppm;
N02, 2.2 ppm; S02 < 0.01 ppm) for 1 day or 1, 4, or 8 weeks (7h/day, 5 days/week, endpoints 12-h post-
DE exposure). Results from the 8 week study are described in Section 7.3.7.1. It should be noted that
control mice were left in a clean room as opposed to undergoing chamber exposure to filtered air.
Significant AHR upon methacholine challenge was observed after 1 and 4 weeks of exposure, and airway
sensitivity (provocative concentration of methacholine causing a 200% increase in Penh) was
significantly increased after 1 week of exposure but not 4 weeks. DE had no effect on total cells in BALF,
but transiently increased expression of IL-4, IL-5, and IL-13 after 1 day of exposure, MDC after 1 week,
and RANTES after 2 and 3 weeks. Eotaxin, TARC, and MCP-1 were elevated, but did not achieve
statistical significance, after short-term (1 day or week) exposure. Statistical power may have been
lacking due to an n of 3. Protein levels of IL-4 and RANTES were significantly elevated after 1 day of
December 2008
6-155
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
DE exposure, respectively. DE had no effect on OVA challenge-induced peribronchial inflammatory or
mucin positive cells. Therefore DE-induced airway hyperreactivity was observed in the absence of
cellular inflammation, similar to the responses described for aerosolized or nebulized DEP by Farraj et al.
(2006a, b) and Hao et al. (2003).
Exposures to an aerosol of soot and iron oxide generated from ethylene (PM2 5, 0.235 mg/m3) were
conducted to test whether the sequence of exposure to ovalbumin aerosol challenge and PM affected the
observed response of OVA sensitized BALB/c mice (Last et al., 2004). Though called PM2 5, the authors
characterized the PM material as ultrafine, 80-110 nm, with the iron oxide crystals often spatially
segregated from the soot (200 (ig/m3 soot, remainder iron oxide, CO < 0.8 ppm, NOx < 0.4 ppm, PAH
below detection). Mice were exposed to PM via chamber inhalation for 2 weeks (4h/day, 3 days/wk)
before or after 4 weeks of OVA inhalation, or simultaneously to PM and OVA for 6 weeks. Among
endpoints (BAL cells, Penh, airway collagen, and goblet cells) only goblet cell counts were significantly
increased with PM exposure in any combination with ovalbumin. There was a trend toward increased
Penh responses with exposure to PM alone or with OVA, particularly when PM exposure immediately
preceded methacholine challenge (after or during OVA challenge). Results from this study are difficult to
interpret due to the varied elapsed times between cessation of PM or OVA treatment and endpoint
determination. The mild responses to PM may be related to the intraperitoneal sensitization protocol used,
reputed to generate a highly allergic mouse in which any additive effects of PM may be obscured by
maximal responses to antigen challenge (Deurloo et al., 2001; Hao et al., 2003).
Pregnancy or in utero exposure may confer susceptibility to PM-induced asthmatic responses.
Exposure of pregnant BALB/c mice to aerosolized ROFA leachate by inhalation or to DEP intranasally
increased asthma susceptibility in their offspring (2008; Hamada et al., 2007). The offspring from dams
exposed for 30 min to 50 mg/ml ROFA 1, 3, or 5 days prior to delivery responded to OVA immunization
and aerosol challenge with AHR and increased antigen-specific IgE and IgGl antibodies. AHR was also
observed in the offspring of dams intranasally instilled with 50 |_ig of DEP or Ti02, or 250 |_ig carbon
black, indicating that the same effect could be demonstrated using relatively "inert" particles. Pregnant
mice were particularly sensitive to exposure to DEP or Ti02 particles, and genetic analysis indicated
differential expression of 80 genes in response to Ti02 on the pregnant background. Thus pregnancy may
enhance responses to PM, and exposure to even relatively inert particles may result in offspring
predisposed to asthma.
Allergic Sensitization
A large number of in vivo animal studies and in vitro studies have demonstrated that particles can
alter the immune response to challenge with specific antigens and suggest that PM may act as an adjuvant
December 2008
6-156
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
to promote allergic sensitization. This phenomenon was introduced in the 2002 Diesel Document, and has
been noted in multiple animal and human studies by the 2004 CD. Adjuvants enhance the immune
response to antigens. Importantly, adjuvants may be major contributors to the development of
inappropriate immune responses. These immune responses, mediated by T helper cells, fall along a
continuum from T helper type 1 (Thl) to T helper type 2 (Th2). Thl responses, characterized by IFN-y,
are inflammatory and in excess can lead to tissue damage. Alternatively, Th2 responses are characterized
by IL-4, IL-5, IL-13, eosinophils, and IgE, and are associated with allergy and asthma. Autoimmune
diseases may be driven by Thl, Th2, or mixed responses, but allergic diseases are Th2 mediated.
Adjuvant activity can be exerted via chemoattraction, cytokines, or enhanced antigen presentation and
costimulation, and may originate via effects on a number of cell types.
It has been suggested that the capacity of particles to enhance allergic sensitization is associated
more strongly with particle number and surface area than particle mass, and several studies comparing
size fractions of the same material show greater adjuvant activity for an equivalent mass dose of smaller
particles (de Haar et al., 2006; Inoue et al., 2005; Nygaard et al., 2004). This is particularly true of inert or
homogeneous materials, such as carbon, polystyrene, and Ti02, which vary little in composition with size
fraction. In studies of ambient PM, however, coarse particles have demonstrated equal and sometimes
greater adjuvancy compared to fine particles (Nygaard et al., 2005; Steerenberg et al., 2004a; Steerenberg
et al., 2005). It is difficult to ascertain the role of particle size in mediating adjuvancy due to a lack of
inhalation studies performed to compare size fractions. The adjuvant effects of ambient PM appear to
depend on composition, which differs among various size fractions and sources, and are associated with
combustion related materials (Steerenberg et al., 2006) and metal content (Gavett et al., 2003).
DEP inhalation during allergen exposure has been shown to augment IgE production and alter
methylation of T helper genes in BALB/c mice (Liu et al., 2008) Animals were exposed to 1280 |_ig/m3
DEP over a 3-week period, 5h per day, concurrent with periodic intranasal sensitization to the common
fungus Aspergillus fumigatus. Gas concentrations were not reported. Total IgE and BALF eosinophils
were elevated with A. fumigatus sensitization and further increased by concomitant DEP exposure.
Greater methylation of the IFN-y promoter was observed following DEP and A. fumigatus exposure (but
not DEP alone) compared to A. fumigatus alone, indicating that combined DEP and allergen exposure
might induce methylation and thus suppress expression of Thl genes. Furthermore, hypomethylation of
the IL-4 promoter was detected after exposure to A. fumigatus and DEP compared with exposure to A.
fumigatus or DEP alone, suggesting pro-allergic Th2 gene activation upon combined exposure to allergen
and DEP. The changes in methylation status of these genes were associated with alterations in IgE levels
in individual animals, indicating that modifications at the genetic level could result in predicted
December 2008
6-157
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
downstream effects. This study shows for the first time that DEP exposure can exert pro-allergic in vivo
effects on the mouse immune system at the epigenetic level.
Resuspended DE particles have been shown to enhance OVA-specific IgGl and IgE in BALB/c
mice exposed via inhalation to doses as low as 200 and 600 (ig/m3, respectively (Whitekus et al., 2002).
Mice were exposed to DEP (200, 600 and 2000 (.ig/m3) for 1-h daily for 10 days, followed by 20 min
exposure to aerosolized OVA. Compared with responses to OVA alone, antibody levels were increased by
all OVA+DEP exposures. Statistical significance was reached for IgGl at all DEP exposure levels,
whereas OVA specific IgE was significantly increased at the 600 and 2000 |_ig/m3 doses and total IgE was
significantly elevated at 2000 (ig/m3. Although strong adjuvant effects were observed, no general markers
of inflammation such as eosinophils, IL-5, GM-CSF, mucin, morphological changes, or eosinophilic
major basic protein (MBP) deposition in the airways were observed in exposed mice. In vitro experiments
using the RAW 264.7 macrophage-like cell line indicated a DEP-induced lipid peroxidation and protein
oxidation, which could be inhibited by a variety of antioxidants. Also observed was a decrease in the
GSH: GSSG ratio and an increase in HO-1 expression, both of which were inhibited only by the thiol
antioxidants NAC and BUC. These same thiol antioxidants were able to completely block DEP-related
increases in IgE and IgGl, as well as lipid peroxides and oxidized proteins recovered from lung lavage
fluids at the 2000 (ig/m3 dose. Thus solid correlations between in vivo and in vitro antioxidant activities
were found, and the reversal of adjuvant effects by antioxidants in vivo clearly indicates a link between
oxidative stress and DEP adjuvant activity. However, the intranasal adjuvant activity of Ottawa dust
(EHC-93) in the same strain of mice was not inhibited by NAC pretreatment (Steerenberg et al., 2004a),
suggesting that disparate pathways may be utilized by different materials to exert immune stimulation.
A toxicogenomic approach to investigate early response mechanisms of DEP adjuvancy was taken
by Stevens et al. (2008). BALB/c mice were chamber exposed to filtered air, 500 or 2000 (ig/m3 PM in
DE for 4 h/day over 5 consecutive days, and intranasally exposed to OVA on each of the first 3 days. In
the low vs. high PM exposures, CO, NO, N02, and S02 were <0.1 vs. 4.3, < 2.5 vs. 9.2, < 0.25 vs. 1.1
and < 0.06 vs. 0.2 ppm, respectively. Lung tissues were assessed for alterations in global gene expression
(n = 4) 4 h after the last DE exposure on day 4. Mice were intranasally challenged with OVA or saline on
day 18 and then with OVA on day 28. Post-challenge results demonstrated mild adjuvancy with antigen
and DE exposure as evidenced by significant increases in eosinophils, neutrophils, lymphocytes, and IL-6
in the BALF. Antibody responses were not significantly affected by DE exposure, although a slight
increase in IgE after high dose exposure was observed. DE alone only increased neutrophils, indicating
the need for combined exposure to DE and antigen in the development of allergic outcomes. Comparison
of low DE (500 |_ig/m3)/OVA versus air/OVA resulted in no significant changes in gene sets associated
with this treatment. Comparison of the high (2000 (.ig/nr3) DE/OVA versus air/O VA, however, showed
December 2008
6-158
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
significant changes in 23 gene sets, including neutrophil homing chemokines, inflammatory cytokines,
numerous interleukins and TNF subtypes, growth and differentiation pathways, and an array of
chemokines.
Interestingly, although the adjuvant effects of DEP are reasonably well established, hardwood
smoke (HWS) exposure only minimally exacerbates indices of allergic airway inflammation in an
OVA-sensitized BALB/c mouse model and does not alter Thl/Th2 cytokine levels (Barrett et al., 2006).
Trend analysis indicated increasing BALF eosinophils with increasing dose of HWS, becoming
significantly elevated at 300 (ig/m3 (CO, 1.6 ± 0.3 ppm; total vapor hydrocarbon, 0.6 ± 0.2 ppm; NOx,
below limit of quantitation), and increasing but not statistically significant OVA-speciflc IgE levels with
HWS up to 1000 (ig/m3.
Summary
Studies conducted since the last review confirm and extend findings of the 2004 CD. Thus PM can
modulate immune reactivity in both humans and animals to promote allergic sensitization and exacerbate
allergic responses. Numerous forms of PM, including inert materials, have been shown to function as
adjuvants, and although studies of relatively homogeneous materials demonstrate greater adjuvancy for
smaller particles, analyses of ambient PM do not. Particle composition is likely more influential than size,
but few if any studies have compared size fractions of well-characterized ambient PM for adjuvant
activity in a direct, controlled fashion via inhalation exposure.
6.3.7. Host Defense
Several toxicological studies were cited in the 2004 PM AQCD that demonstrated increased
susceptibility to infectious agents following exposure to PM. A limited number of new studies have
evaluated the effect of PM on host defense in rodents. Two recent studies have observed an increase in
susceptibility to influenza infection and respiratory syncytial virus in mice. However, one new study
found that wood smoke had no effect on bacterial clearance in rodents.
6.3.7.1. Toxicological Studies
The normal, and very important, role of respiratory immune defense is the detection and/or
destruction of pathogens that enter the lung via inhalation and removal of damaged, transformed
(cancerous), or infected cells. Innate immune defenses of the respiratory tract include mucociliary
clearance, release of toxic antimicrobial proteins into airway surface liquid, and activation of alveolar
macrophages. The innate immune system is the earliest responder to irritation or infection, initiating the
December 2008
6-159
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
normal inflammatory response including the majority of detrimental inflammatory processes discussed.
Activated macrophages and epithelial cells release cytokines and chemokines that can bring into play the
adaptive immune system, which in turn can produce long-lasting pathogen-specific immune responses
critical for resolving and preventing infections.
Several studies included in the 2004 CD demonstrated increased susceptibility to infectious agents
following exposure to various forms of PM, and studies of CAPs exposed aged rats demonstrated
increased S. pneumoniae burdens when a 24-h exposure (65 (.ig/m3) followed infection (Zelikoff et al.,
2003). In another study, exposure to ROFA by intratracheal instillation was found to affect bacterial
clearance (Antonini et al., 2002). Examinations of mechanisms related to PM interference with host
defenses have demonstrated impaired mucociliary clearance and modified macrophage phagocytosis and
chemotaxis. Prolonged exposure to inhaled particles at sufficiently high concentrations can lead to
diminished clearance of PM from the alveolar region of the lung, resulting in the accumulation of retained
particles and an accompanying chronic alveolar inflammation. Diminished clearance of PM may also
increase susceptibility to pulmonary infection by impeding clearance of pathogens. Impaired phagocytosis
by alveolar macrophages may contribute to a decrease in the lung's capacity to deal with increased
particle loads (as occurs during high-pollution episodes) or infections and affect the local and systemic
responses through the release of biologically active compounds (cytokines, ROS, NO, isoprostanes).
Since the last review, several additional studies have reported impairment of pathogen clearance
following intratracheal instillation of PM. In a follow up study Antonini et al. (2004) compared sources of
ROFA in SD rats. Precipitator ROFA induced an inflammatory response and diminished pulmonary
clearance of L. monocytogenes while air heater ROFA had no effect on lung bacterial clearance at the
same dose of 1 mg/lOOg body weight. Precipitator ROFA generated a metal-dependent hydroxyl radical
suggesting that differences in metal composition were a determinant of the immunotoxicity of ROFA.
Viral respiratory infections in early life are associated with increased incidence of childhood
asthma and other pulmonary diseases. DE exposure can enhance the progression of influenza infection
when a virus is introduced following particle exposure. BALB/c mice that were chamber exposed to DE
and subsequently instilled with influenza A/Bangkok/1/79 virus had increased susceptibility to influenza
infection (Ciencewicki et al., 2007). Exposures to two doses of DEP were conducted: 500 (ig/m3 (0.9 ppm
CO, < 0.25 ppmN02, < 2.5 ppm NO, and 0.06 ppm S02) and 2000 (ig/m3 (5.4 ppm CO, 1.13 ppm N02,
10.8 ppm NO, and 0.32 ppm S02). Responses were greater for animals exposed to 500 (ig/m3 DEP than to
2000 |ag/m3. and were associated with a significant increase in IL-6 protein and mRNA expression and
IFN-(3 expression. The authors present the possibility that damage to the epithelium at the higher dose
prevented viral infection and replication. After exposure to 500 |_ig/m3 DEP alone or prior to infection,
December 2008
6-160
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
decreased expression of surfactant proteins (SP) A and D was observed. These proteins are part of the
IFN-independent defense against influenza.
Similarly, Harrod et al. (2003) demonstrated decreased SP-A expression in the lungs following DE
exposure and linked it to increased susceptibility to respiratory syncytial virus (RSV), the most common
cause of respiratory infection in young children. C57BL/6 mice, a relatively RSV-resistant strain, were
exposed via inhalation to DE at a concentration of 30 or 1000 (ig/m3 DPM 6h/day for 7 consecutive days
prior to intratracheal viral inoculation. Gaseous copollutants ranged from 2.0-43.3 ppm for NOx (~ 90%
NO), 0.94-29.0 ppm CO, and 8.3-364.9 ppb S02. Exposure to 30 |_ig/m3 DEP did not induce a
statistically significant increase in BALF cell numbers compared to air-treated, infected animals.
However, distinct consolidated inflammatory infiltrates were observed in the peribronchial regions of
RSV-infected animals exposed to this dose, along with alterations in Clara cell morphology, decreased
CCSP production by these cells, and occasional regional myofibril layer thickening. These changes were
more pronounced in RSV-infected animals exposed to 1000 (ig/m3, and the higher dose also resulted in
significant increases in inflammatory cells, predominantly macrophages, in both uninfected and infected
mice compared to air-exposed controls. Both doses elicited significant levels of TNF-a and IFN-y in the
lungs of infected animals, but decreased levels of SP-A. Consistent with this study's finding of decreased
SP-A and increased viral gene and inflammatory cytokine expression after DE exposure, SP-A" " mice
demonstrate decreased clearance of RSV concordant with increased lung inflammation (LeVine et al.,
1999). Thus, DEP may enhance susceptibility to respiratory viral infections by reducing the expression
and production of SP (Ciencewicki et al., 2007; Harrod et al., 2003), although the contribution of gaseous
copollutants, in some instances concentrated lOOOx, should be considered for both studies. SP are also
essential for clearance of other pathogens, including group B Streptococcus (GBS), Haemophilus
influenzae, and Pseudomonas aeruginosa (LeVine and Whitsett, 2001).
In lung epithelial cells, DEP increased the mRNA expression of intercellular adhesion molecule-1
(ICAM-1), low-density lipoprotein (LDL) and platelet-activating factor (PAF) receptors, which can act as
receptors for viruses or bacteria (Ito et al., 2006b). DEP may therefore enhance the susceptibility to
infection by the upregulation of bacterial and viral invasion sites in the lungs. Expression of the
(3-defensin-2 gene, which is one antimicrobial mechanism of host defense in the airway, was significantly
inhibited by V and not Ni or Fe in airway epithelial cells incubated with aqueous leachate of ROFA
(Klein-Patel et al., 2006).
HWS does not appear to have significant impact on pathogen clearance. Fisher 344 rats, SHR rats,
A/J mice and C57BL/6 mice were exposed to 30-1000 (ig/m3 HWS by whole body inhalation for one
week and 6 months (Reed et al., 2006). Long-term responses are discussed in Sections 7.3.3.2 and 7.3.8.
Concentrations of gases ranged from 229.0-14887.6 mg/m3 for CO, 54.9-139.3 (ig/m3 for ammonia, and
December 2008
6-161
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
177.6-3455.0 (ig/m3 nonmethane volatile organic carbon in these exposures. Bacterial clearance and
inflammation in response to instilled Pseudomonas aeruginosa were unaffected by HWS.
Immunosuppressive Effects of PM
DEP may affect systemic immunity. Decreased thymus weight was observed in female F344 rats
exposed to 300 (ig/m3 DEP for one week by Reed et al. (2004). Concentrations of gases for this dose were
reported to be approximately 16.1 ppm for NO, 0.8 ppm forN02, 9.8 ppm for CO, 115 ppb for S02, and
1416 (ig/m3. Long-term responses are discussed in Section 7.3.8.
6.3.8. Respiratory ED Visits, Hospital Admissions and Physician Visits
The epidemiologic evidence presented in the 2004 PM AQCD of an association between PM and
respiratory hospitalizations and emergency department visits was consistently positive across studies.
Recent studies have provided further support to this relationship, with larger effect estimates observed
among children and older adults. However, effect estimates are clearly heterogeneous, with evidence of
both regional and seasonal differences at play.
Excess risk estimates for hospitalizations or ED visits for all respiratory diseases combined,
reported in studies reviewed in the 2004 PM AQCD fell within the range of approximately 1 to 4% per
10 |ig/m3 increase in PM10. On average, excess risks for asthma were higher than excess risks for COPD
and pneumonia. The limited body of evidence reviewed in the 2004 PM AQCD also reported associations
with ambient fine particles (PM25, PM,) and coarse thoracic particles (PM10_2.5). Excess risk estimates fell
within the range of approximately 2.0 to 6.0% per 10 (.ig/ni3 increases in PM2 5 or PM10_2.5 for both all
respiratory and COPD admissions, whereas larger estimates were reported for asthma admissions. Many
of the associations for respiratory admissions and ED visits and short-term PM2 5 exposure were
statistically significant. The associations with PMi0.2 5 were less precise with fewer reaching statistical
significance (U.S. EPA, 2004). In addition, several studies had reported associations with outpatient
physician visits that suggested larger public health impacts.
Hospital admissions or ED visits for respiratory diseases and ambient concentrations of PM have
been the subject of approximately 80 peer-reviewed research publications since 2002 (see Annex D).
Included among these new publications are several large single-city and multicity studies. The new
studies complement those reviewed in the 2004 PM AQCD and based upon those reults to evaluate effect
modification by season and region as well as the effects of different PM size fractions and components on
admissions or visits for specific diseases of the respiratory system, including asthma, bronchitis and
emphysema, (collectively referred to as COPD), pneumonia, upper respiratory infections, lower
December 2008
6-162
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
respiratory infections and other minor categories. ICD codes (both 9th and 10th revisions) from hospital
admission or discharge records are used to ascertain the outcomes in these studies (Table 6-10). Several
new studies have also evaluated associations between short-term PM exposure and outpatient physician
visits.
Specific design and methodological considerations of the studies reviewed were discussed
previously (Section 4.4.10.2). Like the CVD endpoints discussed, the respiratory endpoints examined in
these studies were also heterogeneous. MCAPS investigators included COPD (ICD-9: 490-492) and
respiratory tract infections (464-466, 480-487) among those 65 years and older in their analyses. SOPHIA
investigators included asthma (ICD-9: 493), wheezing (ICD-9: 786.09), COPD (ICD-9: 491, 492, 496),
lower respiratory infection (466.1, 480-486), and all respiratory diseases combined (ICD-9: 460-466, 477,
480-486, 491-493, 496, 786.09) among all ages. APHEA-2 investigators examined asthma among ages 0-
14 and 15-64, COPD and asthma (ICD-9: 490-496) among older adults and all respiratory diseases
combined among older adults (ICD-9: 460-519). French PSAS investigators examined all respiratory
diseases (ICD-10: J00-J99) and respiratory infection (ICD-10: J10-22) among children, adults and older
adults. Finally, investigators conducting the multicity studies in Australia and New Zealand examined all
respiratory diseases (ICD-10 J00-J99 excluding J95.4-J95.9, R09.1, R09.8); asthma (ICD-10 J45, J46,
J44.8) and pneumonia and acute bronchitis (ICD-10 J12-J17, J18.0, J18.1, J18.8, J18.9, J20, J21) among
children less than 14 years old (scheduled admissions, transfers from other hospitals and admissions
arranged through a general practitioner were excluded). Unless otherwise specified, effect estimates are
presented for an increment of 10 |_ig/m3 increase in PM mass.
Table 6-10. Description of ICD-9 and ICD-10 codes for diseases of the respiratory system.
Description
ICD 9 Codes
ICD 10 Codes
Diseases of the Respiratory
System
460-519
J00-J99
Asthma
493
J 45
COPD and allied conditions
490-496 (asthma, chronic bronchitis, emphysema,
bronchiectasis, extrinsic allergic alveolitis)
Chronic lower respiratory diseases
J40-J47 (bronchitis, emphysema, other COPD, asthma,
status asthmaticus, bronchiectasis)
Acute Respiratory Infections
460-466 (common cold, sinusitis, pharyngitis, tonsillitis,
laryngitis & tracheitis, bronchitis & bronchiolitis)
Acute Upper Respiratory
Infections
J00-J06 (common cold, sinusitis, pharyngitis, tonsillitis,
laryngitis & tracheitis, croup & epiglottitis)
Acute bronchitis and bronchiolitis
466
J20-J22
Pneumonia
480-486
J13-J18
Wheezing
786.09
December 2008
6-163
DRAFT-DO NOT QUOTE OR CITE
-------
6.3.8.1. All Respiratory Diseases
PM10
1 Estimates of the effect of PMi0 on respiratory diseases combined are summarized in Figure 6-9.
2 Information on the PM concentrations during the relevant study periods is found in Table 6-11. Excess
3 risks reported for studies of respiratory hospitalization or ED visit and PM10 reviewed in the 2004 PM
4 AQCD were in the range of approximately 1-4%.
Table 6-11. PM concentrations in studies of respiratory diseases published since 2002.
Pollutant Author
Location
Mean Concentration (|jg/m3)
Upper Percentile: concentrations (|jg/m3)
PMw
Barnett et al. (2005)
7 Cities, Australia, NZ
16.5-20.6
Max: 50.2 - 156.3
Ulirsch et al. (2007)
Idaho
23.2
Max: 183.0
Peel et al. (2005)
Atlanta, GA
27.9
Max: 44.7
Luginaah et al. (2005)
Ontario, Canada
50.6
Max: 349
Slaughter etal. (2005)
Spokane, WA
NR
Max: 41.9 (using 90% of concentrations)
Fung et al. (2005b)
Ontario, Canada
38
Max: 248
Fung et al. (2006)
Vancouver, Canada
13.3
Max: 52.17
Chen et al. (2005b)
Vancouver, Canada
13.3
Max: 52.2
Jaffeet al. (2003)
Cincinnati
43
Max: 90
Cleveland
60.8
Max: 183
Columbus
37.4
Max: 87
Lin et al. (2002b)
Toronto, Canada
30.16
Max: 116.20
Chimonas and Gessner
(2007)
Anchorage, Alaska
27.6
Max: 421
Medina-Ramon et al. (2006)
36 US Cities
15.9-44.0
NR
Yang et al. (2004b)
Vancouver, Canada
13.3
Max: 52.2
Andersen et al. (2007b)
Copenhagen,
Denmark
25/24
75th: 30/99th: 72
Sinclair and Tolsma (2004)
Atlanta, GA
29.03
NR
Chardon et al. (2007)
Paris
23
Max: 97.3
Gordian and Choudhury
(2003)
Anchorage, AK
36.11
Max: 210.0
Jalaludin et al. (2004)
Sydney, Australia
22.8
Max: 44.9
PMis
Barnett et al. (2005)
7 Cities Australia, NZ
8.1-11
Max: 29.3- 122.8
Host et al. (2008)
6 Cities France
13.8-18.8
95th: 25.0-33.0
Peel et al. (2005)
Atlanta, GA
19.2
90th: 32.3
Slaughter etal. (2005)
Spokane, WA
NR
Max: 20.2 (using 90% of concentrations)
Belief al. (2008a)
202 US Counties
NR
NR
Fung et al. (2006)
Vancouver, Canada
7.72
Max: 32
Chen et al. (2005b)
Vancouver, Canada
7.7
Max: 32
Babin et al. (2007)
Washington, DC
"low, never reached code red"
NR
December 2008
6-164
DRAFT-DO NOT QUOTE OR CITE
-------
Pollutant
Author
Location
Mean Concentration (|jg/m3)
Upper Percentile: concentrations (|jg/m3)
Chimonas and Gessner
(2007)
Anchorage, AK
6.1
Max: 69.8
Ito et al. (2007)
New York, NY
All year: 15.1
Warm months (Apr—Sep.): 17.5
Cold months (Oct-March): 15.0
All year: 95th: 32
Warm months (Apr-Sep): 95th: 38
Cold months (Oct-March): 95th: 31
Dominici et al. (2006)
204 US Counties
13.4
75th: 15.2
Yang et al. (2004b)
Vancouver, Canada
7.7
Max: 32.0
Andersen et al. (2007b)
Copenhagen,
Denmark
10
99th: 28
Halonen et al. (2008)
Helsinki, Finland
NR; Median = 9.5
Max: 69.5
Sinclair and Tolsma (2004)
Atlanta, GA
17.62
NR
Chardon et al. (2007)
Paris, France
14.7
75th: 18.2
PMlO-25
Host et al. (2008)
6 Cities France
7.0-11.0
95th: 12.5-21.0
Peel et al. (2005)
Atlanta, GA
9.7
90th: 16.2
Slaughter etal. (2005)
Spokane, WA
NR
NR
Peng et al. (2008)
108 U.S. Counties
NR; Median: 9.8
75th: 15.0
Fung et al. (2006)
Vancouver, Canada
5.6
Max: 27.07
Chen et al. (2005b)
Vancouver, Canada
5.6
Max: 24.6
Lin et al. (2002b)
Toronto, Canada
12.17
Max: 68.00
Yang et al. (2004b)
Vancouver, Canada
7.7
Max: 24.6
Halonen et al. (2008)
Helsinki, Finland
NR; Median: 9.9
Max: 101.4
Sinclair and Tolsma (2004)
Atlanta, GA
9.67
NR
ULTRAFINE
Andersen et al. (2007b)
Copenhagen,
Denmark
Mean particles/cm3: 6847
99th: 19,895
Halonen et al. (2008)
NR: Median particles/cm3: 8,203
Max: 50,990
Children
1 Barnett et al. (2005) used a case-crossover design to study respiratory hospital admissions (ICD-9
2 460-519) of children (age groups 0, 1 to 4, 5 to 14 years) in seven cities in Australia and New Zealand
3 during the study period (1998-2001). In this study, using a 0-1 day average lag, increases in respiratory
4 hospital admissions of 2.3% (95% CI: 1.9-7.3) among children 1-4 years old, 2.5% (95% CI: 0.1-5.1)
5 among children 5-14 years old and 2.0% (95% CI: -0.13 to 4.3) among infants less than 1 year old,
6 per 10 |ig/m3 increase in 24-h average PM10 were observed (2005). Luginaah et al. (2005) did not observe
7 significant increases in respiratory hospitalizations among male or female children in Ontario Canada,
8 while Ulirsch et al. (2007) reported increased admissions for respiratory hospitalizations, ED and urgent
9 care visits combined among children <17 years old.
December 2008
6-165
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
All Ages
In a study of 4 million emergency department visits from 31 hospitals in Atlanta, SOPHIA
investigators reported an excess risk of 1.3% (95% CI: 0.4-2.1) per 10 (.ig/nr1 increase in 24-h average
PMio for ED visits for respiratory causes combined (URI, asthma, pneumonia and COPD) among all ages
in Atlanta during the period August, 1998 to August, 2000 (Peel et al., 2005). Reanalyses with four
additional years of data to assess the robustness of estimates to adjustment for PMi0, 03, CO and N02
found that associations for PMi0 and 03 persisted in two-pollutant models (Tolbert et al., 2007). Luginaah
et al. (2005) studied hospitalizations in Ontario Canada. These authors stratified their results by age and
gender, compared time series to case crossover approaches and presented three single day lags. The
results for all ages combined, stratified by gender and all lags are presented in the figure; however, the
strongest, most significant estimate for PM10 was for adult males (15-64 years old). Additional studies
conducted in Spokane, Washington and Ontario, Canada did not provide evidence of increased hospital
admissions for respiratory diseases among all ages (Fung et al., 2005a; Slaughter et al., 2005) while
Ulirsch et al. (2007) reported a significant positive association among all ages in two Southeast Idaho
cities for hospitalizations, ED and urgent care visits combined. This estimate was robust to adjustment for
gaseous pollutants.
Older Adults
Results from one U.S. and three single city Canadian studies offer somewhat consistent evidence
for the effect of PMi0 on respiratory admissions among older age groups. Ulirsch et al. (2007) found
increases in hospitalizations, ED and urgent care visits combined among this age group in 2 cities of
Southeast Idaho. Two studies in Vancouver report increased admissions for respiratory causes with the
largest effects observed for a 3 day moving average (0-2 days) (Chen et al., 2005b; Fung et al., 2006).
Fung et al. (2005a) observed non-significant increases in admissions with PMi0 among older adults in
Ontario, Canada while the study by Luginaah et al. (2005), which was also conducted in Ontario, did not
provide compelling evidence for an effect that was robust to method selection, although some increases
among males were observed. Finally, a study of hospital admissions for cardiopulmonary conditions
combined among older adults (65+ years) in Allegheny County, PA found a positive association with
PMio at lag 0 (Arena et al., 2006).
In a study in Copenhagen Denmark, which was designed primarily to examine the effects of
ultrafine particles and PM sources, PMi0 was associated with hospitalization for respiratory disease
among older adults greater than 65 years (4.6% [95% CI: 1.5-6.9, per 10 (ig/m3, lag 0-4]) (Andersen et al.,
2007b). This association lost precision but was robust to adjustment for total number concentration (3.8%
December 2008
6-166
DRAFT-DO NOT QUOTE OR CITE
-------
1 95% [CI: 0.8-7.6]). After adjustment for CO the effect estimate was 2.5% (95% CI: 0.4-4.6) and after
2 adjustment for N02 the effect estimate was 2.3% (95% CI: 0.5-4.2) (Andersen et al.. 2007a).
Reference
Location
Lag
USEPA AQCD (2004) Range of Excess Risks
Barnett et al. 2005 7 Cities Australia NZ
Ulirsch et al. 2007 '
2 Cities Idaho
Peel et al. 2005* Atlanta, GA
Luginaah et al 2005 Ontario, Canada
Slaughter et al. 2005* Spokane, WA
Ulirsch et ai. 2007** Idaho
Fung et al. 2005
Fung et al. 2006
Ulirsch et al. 2007*
Chen et al. 2005
Fung et al. 2005
Ontario, Canada
Vancouver, Canada
2 Cities Idaho
Vancouver, Canada
Ontario, Canada
0-1
0-1
0-1
0
0-2
1
2
3
1
2
3
1
2
3
0
0
0
0-1
0-2
0
0-2
0-4
0-6
0
0
0-1
0-2
0
0-1
0-2
PM,0
< 1 y Children
-1-4y
5-14 y
-0-17
| Adults (All Ages) |
Female
— Female
— Female
Male
Male
• Male
- All Ages
. 18-64
-< 65 y
[ Older Adults (65+) |
I I II I I I I I I I I I I I I II I I
5 -2 0 2 4 6 8 10 13 16
Excess Risk Estimate
Figure 6-9. Excess risks estimates per 10 pg/m3 24-h average PM10 concentration for studies of
ED visits and hospitalizations for respiratory diseases. Studies represented in the
figure include all multicity studies. Single city studies conducted in the U.S. or
Canada are also included.
December 2008
6-167
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
**Ulirsch et al. (2007) combines ED visits, hospitalizations and urgent care visits in their analyses.
PM2.5, PM10-2.5 and Other Size Fractions
Studies of PM2 5 and PMi 0-2.5 and hospitalization or ED visits for respiratory diseases that were
conducted since 2002 are summarized in Figure 6-10.
Children
A study of seven cities in Australia and New Zealand reported significantly increased risk of
hospitalization associated with PM25 among children (Barnett et al., 2005)}. Increases in respiratory
hospital admissions of 6.4% (95% CI: 2.7-10.3) among infants less than 1 year and 4.5% (95% CI: 1.9-
7.3) among children 1 to 4 years per 10 (.ig/nr1 increase in PM25 (0-1 day average) were observed (2005).
In contrast to these results, French PSAS investigators report an increase of respiratory hospitalizations
associated with PM10.2.5 of 6.2% (95% CI: 0.4, 12.3, 0-1 day avg.) per 10 (.ig/nr1 increase among children
0-14 years old Host et al. (2008). Non-significant increases of 0.8% or below were observed in
association with 10 (ig/m3 increases in PM2 5 in all age categories (2008). A large effect for PMi0_2.5 (31%
95% CI: -4.7, 80) was also observed in a single city study of children less than 3 years old in Vancouver
(Yang et al., 2004b) but does not appear in the figure because its inclusion caused compression and
reduced the readability of other results presented. These authors also studied the effects of PMi0 and PM2 5
but only reported effect estimates for PMi0.2 5 because PM2 5 was not significantly associated with first
respiratory hospitalization.
Adults and All Ages Combined
During the most recent years of the SOPHIA study (August 1, 1998, through August 31, 2000)
PM2 5, PM10_2.5, ultrafine number count and PM2 5 components (sulfate, acidity, EC, OC, and an index of
water-soluble transition metals) were included in the analyses. Study investigators reported similar
findings as MCAPS investigators for older adults (described below), with a larger increase in ED visits
for respiratory diseases associated with PM2 5 compared to PM10_2.5 (Peel et al., 2005). Using an a priori
lag of 0-2 days, excess risks of 1.6% (95% CI: -0.003 to 3.5) per 10 (ig/m3 increase in 24-h average PM25
and 0.6% (95% CI: -3.6 to 5.1) per 10 (ig/m3 increase in PMio_25 were observed. Weaker, less precise
associations with components were reported and no increase with ultrafine particle count was indicated.
In a recent analysis using data from 1998 through 2002 to compare source apportionment methods, Sarnat
et al. (2008) reported that PM2 5 from mobile sources, PM2 5 from biomass burning and sulfate-rich
secondary PM2 5 were associated with respiratory ED visits and associations were robust to the choice of
December 2008
6-168
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
the method (Sarnat et al., 2008). Excess risks were significant, ranging from approximately 2%-4%,
depending on the method.
French PSAS investigators did not report an association with hospital admissions and PM2 5 or
PMio.2.5 among adults 15-64 years old (Host et al., 2008) In a study of respiratory hospital admission and
ED visits (ICD-9 Codes 460-519) among all ages conducted in Spokane, Washington, no associations
were observed with any size fraction of PM considered (e.g. PMi, PM2 5, PMio_2 5, PMi0) (Slaughter et al.,
2005). However, authors observe that there was a suggestion of greater effect estimates with PM2 5
compared to PMi0.2 .5 (Slaughter et al., 2005). Furthermore, several of the same investigators conducted a
source apportionment analysis using daily fine PM filter samples from the same residential monitor in
Spokane (Schreuder et al., 2006). In this investigation, PM2 5 from vegetative burning in the previous day
(lag 1) was associated with respiratory hospital admissions (1.023 [95% CI: 1.009-1.038] per interquartile
range increase in the source marker).
Older Adults
MCAPS investigators observed largely null findings for PM2 5 and respiratory hospitalizations
(COPD, lower and upper respiratory infections) for the U.S. as a whole but reported heterogeneity in
effect estimates across the country that were explained by regional and seasonal factors (Bell et al.,
2008a). The nationwide excess risk of respiratory admissions with PM2 5 was 0.22% (95% posterior
interval (PI): -0.12 to 0.56, lag 0) (Bell et al., 2008a). The largest increases were observed during the
winter in the Northeast (1.05% [95% PI: 0.29-1.82], lag 0) and the Southeast (1.76% [95% PI: 0.60-2.93],
lag 0). Significant increases in respiratory admissions were also observed at lag 2; an increase of 0.94%
(95% PI: 0.22-1.67, lag 2) was observed for the Southwest and an increase of 0.41 (95%PI: 0.09, 0.74)
was observed for the U.S. as a whole. In a multicity Australian study, Simpson et al. examined the
association between fine particles measured by nephelometry and respiratory hospital admissions (ICD-9
460-519) among older adults (65+ years) and reported significant associations (1.055 [95% CI: 1.008-
1.1045], lag 0-1 day average) from a meta-analysis combining effect estimates from all cities (Simpson et
al., 2005). Simpson et al. (2005) considered results from three statistical models, including standard
GAM, which produced similar results.
In an analysis of PMi0.2.5, MCAPS investigators observed small imprecise increases in respiratory
admissions with 24-h PM10_2 5 concentration (0.33% [95% PI: -0.21 to 0.86, per 10 |ig/nr\ lag 0]) (Peng et
al., 2008), which decreased after adjustment for PM2 5 (0.26% [95% PI: -0.32 to 0.84 per 10 (ig/m3 lag 0]).
Associations with PM2 5 increased (0.7% [95% PI: 0-1.5, lag 0]) or persisted (0.6% [95% PI: -0.2, 1.25,
lag 2]), after adjustment for PM10.2 5, however (excess risks estimated from graphs).
December 2008
6-169
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Location
Lao
US EPA AQCD (2004)
Barnett et al 2005
Host et al. 2008
Peel et al. 2005*
Host et al. 2008
Slaughter et al 2005"
Bell et al. 2008
Host et al 2008
Fung et al (2008
Cher et al. 2005
US EPA AQCD (2004)
Host et al 2008
Peel et a I 2005*
Slaughter et al 2005*
Host et al 2008
Perig et at. 2008
Host et al. 2008
Fling et al. 2006
Chen el al. 2005
*ED Visits
Range of Estimates
Australia NZ
5 French Cities
Atlanta, GA
6 French Cities
Spokane. WA
202 US Counties
6 French Cities
Vancouver. Canada
Vancouver, Canada
Range of Estimates
5 French Cities
Atlanta, GA
Spokane. WA
6 F rench Cities
108 US Counties
6 French Cities
Vancouver. Canada
Vancouver, Canada
0-1
0-1
0-1
0-1
0-2
0-1
1
0
1
2
0
0
0-1
0
0-2
0-4
0-6
0
0-1
0-2
0-1
0-2
1
0-1
0
1
2
0-1
0
0-2
0-4
0-6
0
0-1
0-2
pm25
Adults (All Ages)
Older Adults (65-I-)
Adults (All Ages)
Older Adults (65+) |
I I I I I I I I I I I I I I I I I I I I I
-5
2 4 6 8 10
Excess Risk Estimates
13
16
Figure 6-10. Excess risks estimates per 10 [jg/rn3 increase in 24-h average PM2.5 and PM10-2.5 for
studies of ED visits and hospitalizations for respiratory diseases. Studies
represented in the figure include all muiticity studies. Single city studies conducted in
the U.S. or Canada are also included.
December 2008
6-170
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
French PSAS investigators reported a non-significant increase in hospitalizations for respiratory
diseases (ICD-10 J00-J99) with 24-h average PM10.2.5 among older adults while PM2 5 estimates were
closer to the null (Host et al., 2008). Unlike the previously described MCAPS study (Peng et al., 2008),
however, results from 2-pollutant models were not available in the study of Host et al. (2008). Positive
associations of first hospitalization, overall hospitalizations and readmission for respiratory diseases and
PMio.2.5 were also reported among older adults in Vancouver (Chen et al., 2005b). PMi0.2.5 was associated
with an increase of 15% (95% CI: 4.8-22.8) in overall admissions per 10 (.ig/ni1. Increases associated with
PMio.2.5 were larger for readmissions compared to overall admissions. The association for PM2.5 with
overall admissions was 5.1% (95% CI: -4.9 to 13) and the association with readmissions was not larger. In
this study, effect estimates for PMi0 and PMi0.2.5 lost precision but were robust to adjustment for gaseous
pollutants while the estimate for PM2 5 was null after adjustment for gaseous pollutants (Chen et al.,
2005b). In Vancouver, Fung et al. (2006) also report larger effect estimates for PMi0-2.5 than PM2 5 among
adults 65 years and older (Fung et al., 2006). These authors report increased admissions of 1.8%
(95% CI: -2.5 to 5.8) per 10 (.ig/ni1 increase in PM25 and 3.8% (95% CI: 0-7.6) per 10 (.ig/ni1 increase in
PMi0-2.5 (lag 0-1 day average).
In two analyses of data collected in Copenhagen Denmark between 1999 and 2004, size
distribution and total number concentration of ultrafine, accumulation mode and PM10 source
apportionment were investigated in relation to respiratory hospitalizations (J41-42, J43, J44-46) among
adults greater than 65 years of age (Andersen et al., 2007a; 2007b). Of the ultrafine particle metrics
examined, aged secondary long-range transported particles with a median diameter of 212 nm were
significantly associated with respiratory hospital admissions (1.04 [95% CI: 1.01-1.08], per 495
particles/cm3) (2007b). The authors note that it was difficult to separate the effects of PMi0 and NCa2i2,
which were highly correlated in these data. PM2 5 was not associated with respiratory hospitalizations in
these data. However, PMi0 sources including biomass combustion, secondary inorganic compounds, oil
combustion and crustal were associated with respiratory hospitalizations (RR: 1.040, 95% CI: 1.009,
1.072: RR: 1.050, 95% CI: 1.021, 1.081, RR: 1.035 95% CI: 1.006, 1.065 and RR: 1.054, 95% CI: 1.028,
1.081 per interquartile range respectively) (Andersen et al., 2007b).
Summary
As shown in Figure 6-9, excess risks reported for studies of respiratory hospitalization or ED visit
and PM10 concentration are generally within the range observed in multi-city and large single-city studies
published prior to 2002 and included in the 2004 PM AQCD. The two largest new studies reported
positive findings for the association of PMi0 with respiratory diseases combined in children and among all
ages (Barnett et al., 2005; Peel et al., 2005). Not all single city studies report positive associations,
December 2008
6-171
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
however. Design considerations may explain some of the heterogeneity in results. For example, Ulirsch et
al. (2007) who reported positive findings examined hospitalizations, ED visits and urgent care visits
combined. In general, studies comparing multiple statistical methods report consistency across methods
(Fung et al., 2006; 2005; Simpson et al., 2005; Yang et al., 2004). The heterogeneity in effect estimates in
single city studies may also be related to differences between locations in sources of PM, exposure
patterns and climate conditions. In general, larger associations were reported for children and older adults.
Figure 6-10 shows that effect estimates for PM2 5 and PM10-2.5 from both single city U.S. and
Canadian studies and multicity studies are more heterogeneous and less precise than those for PMi0, with
effect estimates for multicity and large single-city studies generally falling in the range reported in the
2004 PM AQCD. The effects of these size fractions on admissions or visits for specific diseases will be
discussed in the sections that follow. Time lags between exposure and hospital admission or ED visit for
respiratory diseases have been shown to vary by both age, disease and other factors (Forastiere and
Faustini, 2008) and will also be discussed in later sections in relation to specific disease outcomes.
Several additional studies conducted outside the U.S. and Canada reported positive associations of
respiratory hospitalizations among a range of age groups using a variety of lags with PMi0 (Bedeschi et
al., 2007; Chen et al., 2006a; Oftedal et al., 2003) and with PM2 5 (Hinwood et al., 2006; Neuberger et al.,
2004; Vigotti et al., 2007) and BS (Bartzokas et al., 2004). Other studies reported no associations with
PM10 or TSP (Llorca et al., 2005; Vegni and Ros, 2004).
6.3.8.2. Asthma
Results from multicity studies of hospital admissions and ED visits for asthma as well as single city
studies conducted in the U.S. and Canada are summarized in Figure 6-11. Concentrations of PM for the
relevant study period are found in Table 6-12.
Children
SOPHIA investigators (Peel et al., 2005) reported that, for several PM mass indicators, the largest
effect estimate observed using the a priori lag (0-2 d average) was the association of PM10 with pediatric
(2-18 years) asthma ED visits (1.6% [95% CI: -0.2 to 3.4]). Asthma admissions in children were
examined in the Australia / New Zealand multicity study (Barnett et al., 2005). Associations for asthma
hospital admissions with PMi0 and PM2 5 were reported but also did not reach statistical significance.
Lin et al. (2002b) used both time series and case crossover approaches to investigate the influence
of PM on asthma hospitalization in children, 6-12 years old, in Toronto from 1981 to 1993. These authors
report relatively small differences in results obtained through bi-directional case crossover and time series
approaches but indicate that unidirectional case crossover methods may overestimate the relative risks. As
December 2008
6-172
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
shown in Figure 3, single to 7-day average lags were investigated and authors note that estimates
increased and appeared to level off at the longer lags. Effect estimates for PM2 5 are not easily
distinguished from the null but PM10_2.5 is significantly associated with asthma admissions among boys
and among girls. These associations were imprecise but robust to adjustment for gaseous pollutants
among all children combined.
Babin et al. (2007) examined asthma-related ED visits among children 1-17 years old in
Washington DC from 2001 to 2004 in relation to PM2 5, ozone and social economic status and reported an
increase in ED visits with ozone but not with PM2 5 (Babin et al., 2007). PM2 5 was not significantly
associated with pediatric hospitalizations for asthma in Oklahoma (Magas et al., 2007).
Two recent European studies evaluated the effect of PMi0, PM25, ultrafine PM and PMi0 sources
(from source apportionment analysis) with asthma hospitalizations or ED visits (Andersen et al., 2007a;
2007b; Halonen et al., 2008). Anderson et al. (2007) found an association between PMi0 attributed to
vehicle emissions and asthma hospitalizations among children 5-18 years (5.4% 95% CI: 0.57, 22.9 per
10 |ig/m3. 0-5 d avg) in Copenhagen, Denmark (Andersen et al., 2007a). In their study examining PM size
fractions, accumulation mode particles number count was most strongly associated with asthma
admissions (e.g., NCa2i2: 1 08 95% CI: 1.00, 1.17 per 495 particles per cm3, lag 0-5) (Andersen et al.,
2007b).
Halonen et al. (2008) examined the association of various size fractions of PM with ED visits for
asthma among children <15 years, and asthma and COPD combined in adults (15-64 years) and older
adults (65+ years) (Halonen et al., 2008). These authors evaluated lags 0 to 5 and noted a different lag
structure depending on age with children experiencing greater effects at lags 3 to 5 days compared to
adults at lag 0. Aitken, accumulation mode particles and traffic-related PM were significantly and most
strongly associated with asthma visits among children, while no association with PMi0_2.5 was observed.
Sinclair and Tolsma (2004) investigated respiratory ambulatory care visits using ARIES data in
Atlanta, GA (also used by SOPHIA investigators) and health insurance records (Sinclair and Tolsma,
2004). These authors evaluated three 3-day moving average lags (0-2, 2-5 and 6-8 days) and reported
relative risks, with no confidence intervals, for significant results only (not included in Figure 6-11). For
childhood asthma outpatient visits, OHC, PMi0_2.5, PMi0, EC and OC were significantly associated with
ambulatory care visits at lags 0-2 or 2-5 days.
Two recent studies in Anchorage used medical records to examine effects of particle exposure on
pediatric asthma outpatient visits, prescriptions for short-acting inhalers (Chimonas and Gessner, 2007),
and school-administered asthma medication (Gordian and Choudhury, 2003). Chimonas et al. (2007)
examined Medicaid claims for asthma-related and lower respiratory infection visits among children less
than 20 years of age for five years (approximately 25,000 children were enrolled in Medicaid each year
December 2008
6-173
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
between 1999 and 2003). Citing work done in the mid-1980's, the authors describe their city's particles as
arising primarily from natural, geologic sources (PM10), and to a lesser extent from local automotive
emissions (PM25) (Pritchett and Cooper, 1985). Using GEE in a time-series analysis of daily and weekly
effects of particle exposure on health outcomes, the authors found that each 10 |_ig/m3 increase in 24-h
average PMi0 was associated with a 0.6% increase (95% CI: 0.1-1.3)) in outpatient visits for asthma. The
same increase in weekly PMi0 concentration resulted in a 2.1% increase (95% CI: 0.4-3.8)) in asthma
visits, after adjustment for gaseous pollutants. No meaningful associations were observed for PM2 5
(Chimonas and Gessner, 2007).
Gordian and Choudhury (2003) used school nurses' records from 1994-1997, in 12 Anchorage
neighborhood schools to examine the association between various moving averages of PMi0
concentration and asthma medication administration. Using a time-series regression model adjusted for
autocorrelation, they found that the best predictor of amount of medication administration was a 21-day
average of PMi0. The estimated coefficient, given for PMi0 x 100, was 7.25 (p = 0.01).
Jalaludin et al. (2004) described above under respiratory symptom outcomes found that for each
10 (ig/m3 increase in PMi0 (lag 0) there was an associated increased risk of doctor visits for asthma
(9.09% [95% CI: 3.32-15.6]). The strongest association was found for a two-day lag (12.09% 95% CI:
4.87, 19.20), and the effect of the lag 0 exposure was unchanged in multipollutant models with ozone or
N02. The mean PM10 level during the study period of 22.8 |ig/nr\ Another study of house calls in Greater
Paris reported an increased association for both PM25 and PM10 with asthma (4.4% [95% CI: -1.3 to 10.4]
and 2.5% [95% CI: -1.7 to 6.8], respectively, 0-3 day average) (Chardon et al., 2007). The mean levels
were 14.7 and 23 |ig/m3 for PM2 5 and PM10, respectively.
All Ages and Adults
Results from the Atlanta SOPHIA study based on the a priori models examining a 3-day moving
average (lag 0-2 days) revealed no statistically significant associations with asthma (ICD-9 493, 786.09)
among all ages for any of the PM metrics studied (PMi0, PM2.5, PMi0.2.5, Particle Count, PM components)
(Peel et al., 2005). However, the 14-day unconstrained distributed lag model produced an excess risk of
9.9% (95% CI: 6.5, 13.5). The authors note that associations of PM2 5 and OC with asthma tended to be
stronger during the warmer months. Sinclair and Tolsma (2004) report a significant association between
adult outpatient visits for asthma and ultravine particles, but not other PM size fractions.
In New York City, Ito et al. (2007) examined numbers of ED visits for asthma among all ages
(ICD-9 493) in relation to pollution levels from 1999 to 2002 (Ito et al., 2007); several weather models
were evaluated. Although the association with N02 was the strongest, PM2 5 was significantly associated
December 2008
6-174
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
with asthma ED visits in each weather model (strongest during the warm months) and remained
significant after adjustment for 03, N02, CO and S02.
Jaffe et al. (2003) examined the effects of ambient pollutants (PM10, ozone, N02 and S02) during
the summer months (June through August) on the daily number of ED visits for asthma among Medicaid
recipients aged 5 to 34 years from 1991 to 1996 in Cincinnati, Columbus and Cleveland. Lags 1 to 3 were
tested and only statistically significant lags were presented. For all cities combined, the overall effect
estimate for 24-h average PMi0 was 1.0% (95% CI: -1.44 to 3.54 per 10 |ig/m3 increase). The effect
estimate for Cleveland was the only significantly elevated estimate (2.3 95% CI: 0.0 to 4.9 per 10 (.ig/nr1
increase). The authors report results from analyses indicating a possible concentration response for ozone
but no consistent effects for PMi0.
Slaughter et al. (2005) reported no associations with ED visits or hospitalizations for asthma,
among all ages, in Spokane Washington for the PM size fractions studied (PMi, PM2 5, PMi0, PMi0.2.5). An
association with CO, which the authors attribute to combustion related pollution in general, was observed,
however (Slaughter et al., 2005). The PM2 5 air quality index (AQI) was not correlated with asthma ED
visits among military trainees in San Antonio, Texas (Letz and Quinn, 2005). Halonen et al. (2008)
reported that, for older adults, PM2 5 and PMi0.2.5, long range transported and traffic-related particles were
associated with asthma and COPD visits combined. For adults 15-65 years, asthma hospitalizations were
associated with accumulation mode and thoracic coarse particles only. As noted above association with a
longer exposure lag was observed among children than adults. In the U.S., Michaud et al. (2004)reported
an association for asthma and COPD ED visits combined with PM, (lag 1) in Hilo, Hawaii in a study
designed to investigate the effect of volcanic fog (Michaud et al., 2004).
Summary
As shown in Figure 6.11, excess risk estimates for asthma are generally larger than those reported
for diseases of the respiratory system combined. The larger effect estimates for asthma were also observed
in studies reviewed in the 2004 PM AQCD. A variety of factors including the underlying distribution of
individual sensitivity and severity, medication use and other personal behaviors can influence the lag time
observed in observational studies (Forastiere and Faustini, 2008). Still, excess risk estimates for asthma
increase with longer lag times in most studies that examine lag structure (2007a; Andersen et al., 2007b;
Chimonas and Gessner, 2007; Halonen et al., 2008; Lin et al., 2002; Peel et al., 2005). Positive
associations were observed with both PMi0.2.5 and PM2 5 but not consistently across study locations.
Investigators from the only U.S. study (Ito et al., 2007) to observe a positive association for PM2 5 with
childhood asthma hospitalization also reported an association with PMi0.2.5 that was stronger than the
association with PM2 5 in the warm season (Ito et al., 2007).
December 2008
6-175
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Location
Laq
US EPA AQCD (2004)
Range of Estimates
Barnett et al 2005
7 Cities Australia/NZ
0-1
0-1
Jaffe et al 2003*
3 US Cities (Ohio)
3
CincinatU, OH
3
Cleveland. OH
2
Columbus, OH
3
Peel et al 2005*
Atlanta, GA
0-2
Lin et al 2002
Toronto, Canada
1
0-2
0-5
0-6
¦j
0-2
0-5
0-6
Chimonas et al. 2006
Anchorage. Alaska
0
0
Peel et al 2005*
Atlanta, GA
0-2
Slaughter et al 2005*
Spokane. WA
0-13
¦
2
3
Bamett et a I 2005
7 Cities Auslralia/NZ
0-1
0-1
Lin et al 2002
Toronto, Canada
1
0-2
0-5
0-6
1
0-2
0-5
0-6
Babin et al 2007*,
Washington DC
0
0
Chimonas et al. 2006
Anchorage. Alaska
0
0
Ito et al. (2007)*
New York, NY
0-1
0-1
0-1
Peel et al 2005*
Atlanta, GA
0-2
Slaughter et al 2005*
Spokane. WA
1
2
3
Lin et al. 2002
Toronto, Canada
1
0-2
0-5
Peel et al 2005*
Slaughter et a! 2005*
Atlanta, GA
Spokane. WA
0-6
1
0-2
0-5
0-6
0-2
1
2
3
PM^s
PMjo-?5
-1-4 y
- 5-14 y
- 5-34 y
- 2-18 y
Boys 6-12 y
-Girts 6-12 y
Inpatient 0-19 y
Outpatient 0-19 y
Adults (All Ages)
—b-
-f"
y Children
- 5-14 y
Boys 6-12 y
- Boys 6-12 y
Boys 6-12 y
Girls 6-12 y
Girls 6-12 y
¦ Girls 6-12 y
1-17 y
- 5-12 y
Inpatient 0-19 y
— Outpatient 0-19 y
, All year Ad Jts (All Ages)
a Warm
-«—Cool
Boys 6-12 y
-Girls 6-12y
Adults (All Ages)
I I I I I I I I I I I I I I II I I I I I I I I I t I I I I I I I I I I I I I I I I I I
22 -16 -10 -4 0 4 8 12 17 22
Excess Risk Estimate
Figure 6-11. Excess risks estimates per 10 pg/m3 increase in 24-h average PM™, PM25 and PM10.2.5
for studies of asthma ED visits* and hospitalizations. Studies represented in the
figure include all multicity studies. Single city studies conducted in the U.S. or Canada
are also included.
December 2008
6-176
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Most additional single city studies conducted in Europe, South America and Asia, have investigated
the associations of asthma hospitalizations or ED visits with TSP, PM10 and PM2 5 and most have reported
evidence of an association (Arbex et al., 2007; Bell et al., 2008b; Chen et al., 2006a; Erbas et al., 2005;
Galan et al., 2003; Kim et al., 2007b; Ko et al., 2007a; Kuo et al., 2002; Lee et al., 2002; Lee et al.,
2006b; Migliaretti and Cavallo, 2004; Migliaretti et al., 2005; Wong et al., 2002b) while a few have not
(Masjedi et al., 2003; Tsai et al., 2006c; Yang et al., 2007).
6.3.8.3. COPD
Results from multicity studies of hospital admissions and ED visits for COPD as well as single city
studies conducted in the U.S. and Canada are summarized in Figure 6-12. Concentrations of PM for the
relevant study period are found in Table 6-11. In one multicity study using MCAPS data for 204 counties
a significant association of about 1% increase in COPD (ICD-9 490-492) hospitalizations was observed
overall with PM25, with the largest effects at lags 0 and 1. Heterogeneity in effect estimates was observed
across the U.S. with a significant increase of about 4% observed in the Southeast at lag 0. In another study
using Medicare data in 36 U.S. cities (1986 to 1999) short-term exposure to PMi0 was associated with an
increase in COPD hospital admissions (ICD-9 490-496, excluding 493) of 1.47% (95% CI: 0.93-2.01,
lag 1) during the warm season (Medina-Ramon et al., 2006).
In Atlanta, SOPHIA investigators reported a comparably sized effect estimate for COPD (ICD-9
491, 492, 496) and 24-h average PM2 5 (1.5% [95% CI: -3.1-6.3, 0-2 d average]) (Peel et al., 2005). The
association of PM10 with COPD reported by Peel et al. (2005) was 1.8% (95% CI: -0.6-4.3). No
associations were reported for PM10_2.5, ultrafine or PM2 5 components (Peel et al., 2005). Slaughter et al.
(2005) reported no associations between any size fraction of PM in Spokane, Washington (PM10, PM2 5,
PMio_2.5) and COPD (ICD-9 491, 492, 494, 496) (Slaughter et al., 2005). However, in a study conducted
in Vancouver, Canada, Chen et al. (2004) reported significant increases in COPD admissions (ICD-9
490-492, 494, 496) for PM10 (16.5% [95% CI: 6.88-27.02, 0-3 d average]) and PM2 5 (17.1% [95% CI:
4.6-31.0) (Chen et al., 2004). These investigators reported a non-significant increase for PMi0_2 5 (10.0%
[95% CI: -1.2-22.8, 0-3 d average]). The effect estimates for PM metrics were diminished after
adjustment for N02, however.
December 2008
6-177
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Location
Lag
US EPA AQCD (2004)
Zanobetti & Schwartz 2003
Medina-Ramon et al. 2006
Peel et al 2005*
Slaughter et al. 2005*
Chen et al. (2004
Dominici et al. 2006
Peel et al. 2005*
Slaughter et al. 2005*
Chen et al. 2004
Peel et al 2005*
Slaughter et al. 2005*
Chen et al. 2004
Range of Estimates
14 US Cities
36 US Cities
Atlanta, GA
Vancouver, Canada
204 US Counties
Atlanta, GA
Spokane, WA
Vancouver, Canada
Atlanta, GA
Spokane, WA
Vancouver, Canada
0-1
0
1
0
1
0-2 d
0-13 DL
1
2
3
1
2
3
0-2
0
1
2
0-2 DL
0-2
1
2
3
1
2
3
0-2
0-2
1
2
3
1
2
3
0-2
PM,0
-f- Cold
Warm
-Ail Ages
¦ All ages
All ages
I I I M I I II I I I I I I I I II I IIIIIIII I I I I I I I I I I I
-12 -7 -3 1 4 7 11 15 19 23
Excess Risk Estimate
Figure 6-12. Excess risks estimates per 10 |jg/m3 increase in 24-h average PM10, PM2.5 and PM10-2.5
for studies of COPD ED visits* and hospitalizations among older adults (65+ years,
unless other age group is noted). Studies represented in the figure include all
multicity studies. Single city studies conducted in the U.S. or Canada are also
included.
December 2008
6-178
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Summary
With the exception of one study conducted in Spokane Washington (Slaughter et al., 2005),
associations have been consistently observed for PM2 5 and PM10 with COPD in multicity and single city
studies conducted in the U.S. and Canada. Associations with PM10_2.5 are more variable. A study that
examined 7 single day lags in association with pooled COPD and asthma ED visits in Finland reports that
PM2 5, PM10-2.5, traffic sources as well as gaseous pollutants had a more immediate effect in older adults
(lags 0 and 1) compared to the children experiencing asthma (3-5 day lags) (Halonen et al., 2008). Several
single city studies conducted outside of the U.S. or Canada are inconsistent with regard to their findings
for PM and COPD (Agarwal et al., 2006; Hapcioglu et al., 2006; Ko et al., 2007b; Martins et al., 2002;
Masjedi et al., 2003; Tenias et al., 2002; Yang et al., 2007).
6.3.8.4. Pneumonia and Respiratory Infections
Results from multicity studies of hospital admissions and ED visits for respiratory infection as well
as single city studies conducted in the U.S. and Canada are summarized in Figure 6-13. Concentrations of
PM for the relevant study period are found in Table 6-11.
Children
In the study of 7 cities in Australia and New Zealand, associations of PM2 5 with pneumonia and
acute bronchitis were observed among infants less than one year old (4.54% [95% CI: 0.00-9.20]) and
children 1-4 years old (6.44% [95% CI: 0.26-12.85]) (Barnett et al., 2005).
In a study of inpatient and outpatient visits for lower respiratory tract infections among children in
Anchorage, Alaska, no significant associations were observed (Chimonas and Gessner, 2007). In contrast,
Lin et al. (2005) observed associations of respiratory infections (ICD-9 464, 466, 480-487) with PMi0.2.5
and PM10 that persisted after adjustment for gaseous pollutants among children less than 15 years old (Lin
et al., 2005). Significant single pollutant excess risks ranged from 14%-35% per 10 (ig/m3, with the
largest effect estimates with PM10.2 5 (results not included in the table because they were imprecise and
other results were compressed). PM2 5 was not associated with respiratory infections in these data.
All Ages
SOPHIA investigators examined ED visits for upper respiratory tract infections (URI) (ICD-9
460-466, 477) and pneumonia (ICD-9 480-486). An excess risk of 1.4% (95% CI: 0.4, 2.5 per 10 (ig/m3,
lag 0-2 d average) for PM10 was associated with URI. With the exception of a small increase in risk for
OC of 2.8% (95% CI: 0.4, 5.3, per 2 (ig/m3, 0-2 d average) with pneumonia, Peel et al. (2005) reported no
association with other PM size fractions or components evaluated (Peel et al., 2005). However, Sinclair
December 2008
6-179
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
and Tolsma, who also used ARIES data for their analysis found significant associations with outpatient
visits for LRI, generally at a 3-5 day moving average lag, with PM10.2.5, PM10, EC, OC, and PM2 5 water
soluble metals (only significant results were reported) (Sinclair and Tolsma, 2004). No associations with
respiratory infections for any size fractions were observed among all ages in a study conducted in
Spokane, Washington (Slaughter et al., 2005).
Older Adults
In a multicity study of older adults (65+ years) Medina-Ramon et al. (2006) examined hospital
admissions for pneumonia (ICD-9 480-487) in 36 U.S. cities in relation to 24-h average PMi0
concentration (Medina-Ramon et al., 2006). An increase in pneumonia admissions of 0.84% (95% CI:
0.50, 1.19, per 10 (ig/m3, lag 0) was reported by these investigators during the warm season. Cold season
associations were weaker (0.30% 95% CI: 0.07, 0.53, per 10 (ig/m3, lag 0) as were lag 1 associations.
Dominici et al. (2006) investigated hospital admissions for all respiratory infections including pneumonia
(ICD-9 464-466, 480-487) among older adults in 204 urban U.S. counties in relation to PM2 5 and
reported a significant increased risk only at lag 2 (Dominici et al., 2006). Heterogeneity in effect
estimates were observed across the U.S. with an association close to the null value reported in the
Northeast (Dominici et al., 2006). In Boston, excess risks of pneumonia hospitalization in association
with PM2 5, BC, and CO were observed among older adults (Zanobetti and Schwartz, 2006). A measure of
non-traffic PM, the residuals from the regression of PM25 on BC, was not associated with pneumonia
hospitalization in these data. In a study of 4 cities in Australia, a statistically significant association of
pneumonia and acute bronchitis with N02 but not PM2 5 was observed among older adults (Simpson et al.,
2005).
Summary
Most of the large single-city and multicity studies of hospitalization for respiratory infection among
the adults, especially those over 65 years, consistently report excess risks associated PM10 and PM2 5.
Only the SOPHIA study of adults examined the association of respiratory infections with PM10.2 5;
investigators reported an imprecise increase in ED visits that was not readily distinguishable from the null
value (Peel et al., 2005). In a study of COPD and respiratory infections combined, MCAPS investigators
did not observe an association with PMi0.2.5 after adjustment forPM2 5 (Peng et al., 2008). The two studies
of children conducted in the U.S. or Canada were conflicting with regard to PMi0.2 5 (Chimonas and
Gessner, 2007; Lin et al., 2005).
December 2008
6-180
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Location
Barnett et al 2005 7 Cities Australia NZ
Medina-Ramon 2006 36 US Cities
Peel et al. 2005* Atlanta, GA
Dominici et al. 2006 204 Urban Counties
Peel et al. 2005* Atlanta, GA
Peel et al. 2005* Atlanta, GA
Lag
0-1
0-1
0
1
0
1
0-2
0-13 DL
0-2
0-13 DL
Barnett et al 2005 7 Cities Australia NZ 0-1
0
1
2
0-2 DL
0-2
0-2
0-2
PMI
< 1 y ¦
4 y
Pneumonia, Cold (*¦
All Ages
PMo
<1
1-4 y
PM,
Pneumonia, Acute Bronchitis
" Pneumonia, Warm
¦ Upper Respiratory Infection
Upper Respiratory Infection
Pneumonia
Pneumonia, Acute Bronchitis
¦ Respiratory Tract Infection
¦ Respiratory Tract Infection
.1 Respiratory Tract Infection
i— Respiratory Tract Infection
• Upper Respiratory Infection
¦ Pneumonia
-Pneumonia
(All ages)
"Upper Respiratory Infection
I I I I I I I I I I I I I I I I I I I
-5 -2 0 2 4 6 8 10 12 14
Excess Risk Estimate
Figure 6-13. Excess risks estimates per 10 |jg/m3 increase in 24-h average PM10, PM2.5 and PM10-2.5
for studies of respiratory infection ED visits* and hospitalizations. Studies
represented in the figure include all multicity studies. Single city studies conducted in
the U.S. are also included.
December 2008
6-181
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Several other single city studies conducted outside the U.S. and Canada reported associations
between PM and hospitalization or ED visits for respiratory infections (Cheng et al., 2007; Hinwood et
al., 2006; Hwang and Chan, 2002; Nascimento et al., 2006).
6.3.8.5. Copollutant Models
Some studies have investigated potential confounding by copollutants through the application of
2-pollutant models. In the MCAPS study the PMi0.2.5 effect of respiratory admissions was not robust to
adjustment for PM2 5 (Peng et al., 2008). In studies of respiratory diseases combined, effect estimates for
PMio were robust to adjustment for gases in several studies (Tolbert et al., 2007; Ulirsch et al., 2007). The
PMio.2.5 effect observed in Vancouver was robust to adjustment for gases while the PM2 5 effect was not
(Chen et al., 2005b). In studies conducted in Copenhagen, PMi0 associations with respiratory disease did
not change in models also containing total number concentration, N02 and ozone (Anderson and Bogdan,
2007). The associations of PM2 5 and PMi0 with asthma were not diminished in 2 pollutant models with
gases (Anderson and Bogdan, 2007; Chimonas and Gessner, 2007; Ito et al., 2007). However, the COPD
association reported by Chen et al. (2004b) in Vancouver was diminished when N02 was included in the
model. Lin et al. (2005) reported that associations between PMi0 and PMi0.2 5 and respiratory infection
remained after adjustment for gases. Studies attempting to distinguish the independent effect of sources
rather than individual highly correlated pollutants have observed effects for vegetative burning (Schreuder
et al., 2006) and traffic sources (Anderson and Bogdan, 2007; Sarnat et al., 2008).
Table 6-12. Characterization of ambient PM concentrations from studies of hospitalization or ED
visits for respiratory diseases
Description
ICD 9 codes
ICD 10 Codes
Diseases of the Respiratory System
460-519
J00-J99
Asthma
493
J45
COPD and allied conditions
490-496 (asthma, chronic bronchitis, emphysema,
bronchiectasis, extrinsic allergic alveolitis)
Chronic lower respiratory diseases
J40-J47 (bronchitis, emphysema, other COPD, asthma,
status asthmaticus, bronchiectasis)
Acute Respiratory Infections
460-466 (common cold, sinusitis, pharyngitis,
tonsillitis, laryngitis & tracheitis, bronchitis & bronchiolitis)
Acute Upper Respiratory Infections
J00-J06 (common cold, sinusitis, pharyngitis, tonsillitis, I
aryngitis & tracheitis, croup & epiglottitis)
Acute bronchitis and bronchiolitis
466
J20-J22
Pneumonia
480-486
J13-J18
Wheezing
786.09
December 2008
6-182
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
6.3.9. Summary and Causal Determinations
6.3.9.1. PM10
Causal Determination
Overall, evidence from epidemiologic studies of respiratory symptoms and medication use provide
consistent evidence for an association with ambient concentrations of PMi0 among asthmatic children.
Similarly, new evidence of ED visits and hospital admissions builds upon the positive and statistically
significant evidence presented in the 2004 PM AQCD to support a consistent association with ambient
concentrations of PMi0, especially for older adults. Despite the lack of human clinical or toxicological
studies, the consistent evidence from epidemiologic studies alone is sufficient to conclude that a
causal relationship is likely to exist between relevant PM10 exposures and short-term respiratory
morbidity.
Respiratory Symptoms and Medication Use
Epidemiologic Studies: The 2004 PM AQCD concluded that the associations for PMi0 with
respiratory symptoms in asthmatics tended to be positive, although they were somewhat less consistent
than PMio effects on lung function. Most studies showed increases in cough, phlegm, difficulty breathing,
and bronchodilator use, although these increases were generally not statistically significant for PMi0. New
studies of respiratory symptoms and medication use provide further evidence of a consistent association
with PMio among asthmatic children (Figure 6-5) but less consistent evidence among asthmatic adults
(Figure 6-7). There was no evidence to suggest an association between respiratory symptoms or
medication use and PM10 among healthy individuals.
Pulmonary Function
Epidemiologic Studies: The peak flow analysis results for asthmatics reported in the 2004 PM
AQCD showed small decrements for PM10. The effect estimates for morning PEF lagged one day showed
decreases, but the majority of the studies were not statistically significant. Several more recent studies
have reported inconsistent results for FEV, and PEF and PM10.
Pulmonary Inflammation
Epidemiologic Studies: There were no epidemiologic studies of pulmonary inflammation
described in the 2004 PMAQCD. Only one recent study looked at the association ofPMio levels and
eNO (Jansen et al., 2005) and reported a positive effect.
December 2008
6-183
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Hospital Admissions and ED Visits
Epidemiologic Studies: Most associations between PMio and hospital admissions for all
respiratory causes reviewed in the 2004 PM AQCD disease were positive and statistically significant.
Studies published since 2002 provide further evidence of a consistent association of PM10 with respiratory
ED visits or hospitalizations within the epidemiologic literature (Barnett et al., 2005; Chen et al., 2005b;
Fung et al., 2006; Peel et al., 2005; Ulirsch et al., 2007). Effect estimates were generally larger among
children and older adults compared to adults or all ages combined. A large U.S. study (Peel et al., 2005)
and a multicity study (Barnett et al., 2005) published since 2002 provide additional evidence of an
association between PMi0 and asthma among children and adults. Also, increases in COPD and
respiratory infection admissions or ED visits with both PMi0 have been consistently observed in large,
recent studies (Barnett et al., 2005; Medina-Ramon et al., 2006; Peel et al., 2005). The few studies that
adjusted PMi0 estimates for the effect of gases in two pollutant models (Anderson and Bogdan, 2007; Lin
et al., 2005; Tolbert et al., 2007; Ulirsch et al., 2007) reported that the PMi0 appeared to be robust. Given
the limitations in distinguishing the independent effects of highly correlated pollutants, these findings
support the conclusion from the 2004 PM AQCD that PMi0, alone or in combination with covarying
pollutants is associated with hospital admissions or ED visits for respiratory causes.
6.3.9.2. PM10-2.5
Causal Determination
A recent analysis of MCAPS data provides new evidence that PMi0.2.5 may not be associated with
respiratory hospital admissions among older adults. However, several epidemiologic studies observed
increases in ED visits and hospital admissions for respiratory causes with acute exposure to PMi0.2.5 in
areas where the mean annual concentrations ranged from 5.6 to 12.2 |ig/nr\ Although these associations
are most consistent in children, increased admissions with PMi0.2.5 concentration have also been observed
in older adults. Results of human clinical studies were mixed with one showing no effect of PM10_2 5 on
respiratory symptoms or pulmonary function in healthy or asthmatic adults and two showing pulmonary
inflammation in healthy adults exposed to PMi0.2.5. This evidence is Suggestive of a Causal
relationship between relevant PM10-2.5 exposures and short-term respiratory outcomes.
Respiratory Symptoms and Medication Use
Epidemiologic Studies: The 2004 PM AQCD presented the results from one study of respiratory
symptoms and thoracic coarse particles (Schwartz and Neas, 2000), which found a statistically significant
December 2008
6-184
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
association with cough. Two new studies examined this relationship and found no association with lower
respiratory symptoms (Mar et al., 2004) or wheeze or medication use (von Klot et al., 2002).
Human Clinical Studies: There were no human clinical studies presented in the 2004 PM AQCD
that evaluated the effect of coarse PM on respiratory symptoms. In the only human clinical study to
examine respiratory symptoms following exposure to thoracic particles, Gong et al. (2004b) found no
effect of exposure to coarse CAPs in healthy or asthmatic adults.
Pulmonary Function
Epidemiologic Studies: The 2004 PM AQCD described two studies that used PMi0.2.5 as a coarse
fraction particulate measure. Tiitanen et al. found that one day lag of PMi0.2.5 was related to morning
PEF, but there was no effect on evening PEF. Neas et al. (1999) found no effects of PM10-2.5 on PEF. There
were no new studies examining the relationship between thoracic coarse particles and pulmonary
function.
Human Clinical Studies: There were no human clinical studies presented in the 2004 PM AQCD
that evaluated the effect of coarse PM on pulmonary function. One new human clinical study reported no
change in lung function in healthy or asthmatic adults following a controlled exposure to coarse CAPs
(2004b).
Pulmonary Inflammation
Human Clinical Studies: The 2004 PM AQCD did not include any human clinical studies that
assessed pulmonary inflammation following controlled exposure to coarse PM. In two new studies,
evidence is presented of a coarse PM-induced increase in neutrophils in BAL fluid and induced sputum in
healthy adults, with additional evidence of alveolar macrophage activation associated with biological
components of coarse PM (Alexis et al., 2006; Samet et al., 2007).
Pulmonary Injury
Toxicological Studies: The relationship between exposure to thoracic coarse PM and markers of
pulmonary injury has not been evaluated in inhalation studies. However an important recent series of
studies has evaluated the relative toxicity of PM size fractions by measuring markers of injury in BALF
of mice following intratracheal instillation or aspiration of ambient PM from a variety of European and
U.S. cities. Markers of inflammation in BALF were also evaluated. Thoracic coarse PM showed greater
potency compared with the other size fractions. Using a similar approach, Gilmour et al. (2004a)
evaluated the relative toxicity of Montana and Western Kentucky coal fly ash of differing size fractions.
December 2008
6-185
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
In contrast to the results above, PM10_2,5 fractions from Montana and Western Kentucky coal fly ash were
less potent than smaller size fractions (fine and ultrafine) in terms of markers of injury and inflammation.
Hospital Admissions and ED Visits
Epidemiologic Studies: Among the key research questions when the 2004 PM AQCD was
published was the relationship between thoracic coarse particles and health outcomes. A number of recent
reports have shown significant associations between respiratory ED visits or hospitalization and acute
exposure to PM10.2.5. The French PSAS program found excess risks for PM10-2.5 and all respiratory
diseases in all age groups, with the strongest relationships observed among children (Host et al., 2008).
Other associations with hospitalizations include those in Vancouver for respiratory illness in children < 3
years of age (Yang et al., 2004b), COPD in the elderly, (Chen et al., 2004b) and respiratory illness in the
elderly (Chen et al., 2005b). Associations were also reported with hospitalization for asthma in children
(Lin et al., 2002b) and respiratory illness in children (Lin et al., 2005) in Toronto. These associations with
hospital admissions for respiratory disease in Canada were observed for PMi 0-2.5 in both time-series and
case-crossover analyses, and the associations remained significant with adjustment for gaseous
copollutants in four of the five studies (except Chen et al., 2005b).
By contrast, MCAPS investigators found that respiratory diseases combined (COPD, upper and
lower respiratory tract infections) were not associated with PM10-2.5 among older adults after adjustment
for PM2 5 nor did they observe heterogeneity in effect estimates across the U.S. (Peng et al., 2008).
Several single city studies support these findings. Slaughter et al. (2005) observed no significant
associations between PM10_2 5 and hospitals admissions or emergency room visits in Spokane, WA for all
ages taken together. Peel et al. (2005) reported no significant associations between PM10.2.5 and
respiratory emergency department visits in Atlanta; however in another Atlanta study, significant
associations were reported between acute PM,n_2 5 exposure and outpatient medical visits for several
respiratory conditions (Sinclair and Tolsma, 2004).
Overall, these studies provide the most consistent evidence for associations between acute PMi0_2.5
exposure and respiratory morbidity among children, and less consistent evidence among adults, given the
recent MCAPS results. Locations with positive findings for PMi0_2.5 and respiratory morbidity reported
mean concentrations range from 5.6 to 12.2 (ig/m3, and maximum concentrations from 24.6 to 68 (ig/m3.
6.3.9.3. PM2.5
Causal Determination
Adverse associations between PM2 5 and hospitalizations and ED visits for respiratory diseases
(e.g., COPD and respiratory infections) have been consistently observed among older adults while the
December 2008
6-186
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
associations of asthma hospitalizations and ED visits with PM2 5 are more heterogeneous. Epidemiologic
studies of asthmatic children have observed increases in respiratory symptoms and asthma medication use
associated with higher PM2 5 concentrations. Additionally, a 10 |ig/m3 increase in PM2 5 is associated with
a decrease in FEVi ranging from 1-3.4% and an increase in eNO ranging from 0.46 to 6.99 ppb in
asthmatic children. Human clinical studies provide coherence and biological plausibility for this
conclusion in that new studies in adults have demonstrated increased markers of pulmonary inflammation
following DE and other traffic-related exposures, oxidative responses to DE and woodsmoke, and
exacerbations of allergic responses and allergic sensitization following exposure to DE particles.
Numerous toxicological studies demonstrating a wide range of responses provide biological plausibility
for the associations between PM2 5 and respiratory morbidity observed in epidemiologic studies. Altered
pulmonary function, mild pulmonary inflammation and injury, oxidative responses, AHR in allergic and
non-allergic animals, exacerbations of allergic responses and increased susceptibility to infections were
observed in a large number of studies involving exposure to CAPs, DE, other traffic-related PM and
woodsmoke. Therefore, the evidence is sufficient to conclude that a causal relationship is likely to
exist between relevant PM2.5 exposures and short-term respiratory morbidity.
Respiratory Symptoms and Medication Use
Epidemiologic Studies: The 2004 PM AQCD reported the results of two studies examining
respiratory symptoms and PM2 5. A review of this limited evidence revealed slightly larger effects for
PM2 5 than for PM10. New studies of respiratory symptoms and medication use provide further evidence of
an association with PM2 5 among asthmatic children (Figure 6-6) but less consistent evidence among
asthmatic adults (Figure 6-7). There was no evidence to suggest an association between respiratory
symptoms and medication use with PM2 5 among healthy individuals. Delfino et al. (2003a) looked at the
EC and OC components of PM and found positive associations with asthma symptoms.
Human Clinical Studies: Studies cited in the 2004 PM AQCD found no effect of PM2 5 CAPs on
respiratory symptoms in healthy adults, although several studies did observe an increase in upper
respiratory symptoms (e.g., nasal irritation) following exposure to DE. One new study found an increase
in respiratory symptoms following exposure to urban traffic PM2 5 (Larsson et al., 2007); however, no
new human clinical studies have evaluated the effect of diesel exposure on respiratory symptoms. Two
new studies found that neither fine zinc oxide, nor fine CAPs caused respiratory symptoms in healthy
adults, or older adults with or without COPD (Beckett et al., 2005; Gong et al., 2004a).
December 2008
6-187
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Pulmonary Function
Epidemiologic studies: The peak flow analysis results for asthmatics reported in the 2004 PM
AQCD tended to show small decrements for PM2 5. More recent studies of pulmonary function and PM25
have yielded somewhat inconsistent results, though the majority of studies have found an association
between PM2 5 concentration and FEV,. PEF, and/or MMEF. Among asthmatic children, a 10 |ig/nr'
increase in PM25 was associated with a decrease in FEVi ranging from 1-3.4%.
Human Clinical Studies: Very little human clinical evidence is presented in the 2004 PM AQCD
of a relationship between fine particulate exposure and changes in pulmonary function. Although one
study reported a significant decrement in thoracic gas volume in healthy adults following exposure to fine
CAPs (Petrovic et al., 2000), several additional studies found no effect of PM2 5 on spirometry or airway
resistance. Two new studies have reported decreases in arterial oxygen saturation following exposure to
PM2 5 CAPs with more pronounced effects observed in healthy adults than in asthmatics or older adults
with COPD (Gong et al., 2004; 2005). In one of these studies, healthy older adults were also reported to
experience a decrease in maximal mid-expiratory flow following exposure to PM2 5 CAPs.Some human
clinical evidence of diesel-induced decrements in lung function was presented in the 2004 PM AQCD.
However, this finding was not consistent across studies. One new study found that exposure to urban
traffic PM did not affect pulmonary function. No human clinical studies have evaluated the effect of
diesel exposure on pulmonary function since the publication of the 2004 PM AQCD.
Toxicological Studies: The 2004 PM AQCD reported changes in respiratory rate and tidal volume
in two out of three studies involving short-term inhalation exposure to CAPs. Furthermore, AHR was
observed in 4 studies of mice, healthy rats or SH rats, exposed to ROFA by inhalation or intratracheal
instillation. Since the last review, SH rats and rats with monocrotaline-induced pulmonary hypertension
demonstrated decreased respiratory rates (Lei et al., 2004b; Nadziejko et al., 2002), altered inspiratory
and expiratory times (Kodavanti et al., 2005), increased tidal volumes (Lei et al., 2004b) and increased
AHR (Lei et al., 2004b) following multi-day exposure to CAPs from a variety of locations. AHR was
observed following 1 week exposure to DE in one mouse strain but not another (Li et al., 2007).
Vagal-mediated pathways were found to be involved in the increased respiratory mean volume observed
in rats exposed to DEP by intratracheal instillation (McQueen et al., 2007).
Pulmonary Inflammation
Epidemiologic Studies: There were no epidemiologic studies of pulmonary inflammation
described in the 2004 PM AQCD. Several recent studies examined PM2 5 and eNO. All of the studies
reported statistically significant, positive effect estimates, though there was inconsistency in the lag times
and use of medication. Among asthmatic children, a 10 (ig/m3 increase in PM2 5 was associated with an
December 2008
6-188
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
increase in eNO ranging from 0.46 ppb to 6.99 ppb. One more recent study (Delfino et al., 2006) found a
positive association between EC concentrations and eNO.
Human Clinical Studies: The 2004 PM AQCD presented the results of one study that reported an
increase in airway neutrophils following exposure to PM2 5 CAPs (Ghio et al., 2000). One new study
found that these effects could be largely attributed to the content of sulfate, iron, and selenium in the
soluble fraction of the PM (Huang et al., 2003c). In addition, the 2004 PM AQCD presents evidence from
human clinical studies of an increase in pulmonary inflammation following exposure to DE. This is
consistent with the observations of several new studies reporting traffic or diesel-induced increases in
markers of inflammation (e.g., neutrophils and IL-8) in airway lavage fluid from healthy adults. No
human clinical studies involving controlled exposures to woodsmoke were presented in the 2004 PM
AQCD. One new study observed an increase in eNO following exposure to woodsmoke in healthy adults.
Toxicological Studies: The 2004 PM AQCD reported that CAPs exposure in rats and dogs at
concentrations of 100-1000 (ig/m3 for 1-6 h/d and 1-3 d generally resulted in minimal to mild
inflammation in healthy animals. More recent studies demonstrate pulmonary inflammation in all three
studies involving multi-day exposure of healthy rats to CAPs from different locations (Godleski et al.,
2002; Rhoden et al., 2004; Smith et al., 2003). No pulmonary inflammation was seen in one study of
WKY and SH rats exposed to CAPs for 4 h and analyzed 1-3 h later (Kodavanti et al., 2005). However,
pulmonary inflammation was seen following a multiday exposure to CAPs in WKY but not SH rats
(Kodavanti et al., 2005). Other investigators using SH rats demonstrated pulmonary inflammation
following multiday CAPs exposure in one study (Cassee et al., 2005) but not another (Kooter et al.,
2006). In the rat monocrotaline model of pulmonary hypertension, multi-day exposures to CAPs resulted
in mild pulmonary inflammation (Lei et al., 2004b). Few studies involving exposure to traffic-related
PM2 5 were discussed in the 2004 PM AQCD. Since then, inflammation was observed in healthy rats
exposed for 20 h, but not 6 h, to ambient air from a high traffic site (Pereira et al., 2007). In contrast, no
inflammation was observed in old rats exposed for 6 h/d for 1-3 day to on-road highway aerosols (Elder et
al., 2004a). The 2004 PM AQCD referred to the 2002 EPA Diesel Document which reported pulmonary
inflammation following short-term inhalation exposure to low levels of DE. Since then, pulmonary
inflammation was demonstrated in three out of five studies in healthy mice and rats exposed by short-term
inhalation to DE. In these three studies, positive effects were seen at concentrations of DEP 100 |_ig/m3
and higher (Harrod et al., 2003; Li et al., 2007; Witten et al., 2005; Wong et al., 2003). No attempt was
made in these studies to determine whether responses were due to PM components or gaseous
components. However, intratracheal instillation of PM from DE resulted in an inflammatory response in
healthy rats (McQueen et al., 2007). This response involved vagally-mediated pathways. No
inflammatory response to gasoline exhaust was seen in one study (Campen et al., 2006) while in another
December 2008
6-189
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
study, the inflammatory responses to gasoline exhaust were due to gaseous components (Sureshkumar et
al., 2005).
The 2004 PM AQCD reported pulmonary inflammation in healthy rats and mice exposed by
inhalation to fine particles of Ti02 and carbon black at extremely high concentrations. Since then, one
study demonstrated an inflammatory response in healthy rats at a much lower concentration of carbon
black (1400 (.ig/ni') (Gilmour et al., 2004c).
Oxidative Responses
Human Clinical Studies: One study cited in the 2004 PM AQCD reported increases in alveolar CO
following controlled exposure to diesel (Nightingale et al., 2000), while another (Blomberg et al., 1998)
found a diesel-induced increase in ascorbic acid concentrations in nasal lavage immediately following
exposure. New studies have provided additional evidence in support of a pulmonary oxidative response to
DE including induction of redox-sensitive transcription factors, as well as increased urate and GSH
concentrations in nasal lavage. No human clinical studies involving controlled exposures to wood smoke
were presented in the 2004 PM AQCD. However, a recent human clinical study found an increase in the
levels of malondialdehyde in breath condensate of healthy adults following exposure to wood smoke
(Barregard et al., 2008).
Toxicological Studies: The 2004 PM AQCD reported one study which provided evidence that
ROS were involved in PM2 5-mediated effects. New studies demonstrate that CAPs exposure in healthy
rats resulted in increased lung ROS measured by chemiluminescence; increased oxidation products of
lung lipids and proteins; and increased MnSOD and catalase (Gurgueira et al., 2002; Rhoden et al., 2004).
Lipid peroxidation was found in healthy rats exposed for 20 h to ambient PM from a high traffic site
(Pereira et al., 2007) and in allergic mice exposed to DE (Whitekus et al., 2002). Increased HO-1 was
observed in SH rats exposed to CAPs (Kooter et al., 2006) and in healthy mice exposed to DE (Li et al.,
2007). Further evidence for involvement of oxidative responses is provided by studies using pretreatment
with the thiol antioxidant NAC to inhibit the PM-mediated responses (Rhoden et al., 2004; Whitekus et
al., 2002) (Li et al., 2007).
Pulmonary Injury
Toxicological Studies: The 2004 PM AQCD reported pulmonary injury in healthy and
compromised animals following inhalation or intratracheal instillation of ROFA or other metal-containing
PM. Intratracheal instillation of diesel PM also resulted in injury. Mild increases in markers of pulmonary
injury were noted in several studies involving inhalation exposure to CAPs. More recent studies
demonstrated mild injury accompanying the mild inflammatory responses to CAPs in SH rats (Cassee et
December 2008
6-190
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
al., 2005; Kodavanti et al., 2005) and in rats with monocrotaline-induced pulmonary hypertension (Lei et
al., 2004b). In addition, CAPs exposure resulted in a mild lung edema in two studies involving healthy
rats (Gurgueira et al., 2002; Rhoden et al., 2004). Exposure to DE resulted in plasma extravasation in
healthy rats (Witten et al., 2005; Wong et al., 2003).
Allergic Responses
Human Clinical Studies: Exposure to DE particles in controlled human exposures studies has been
shown to increase the allergic response among previously sensitized atopic subjects. In addition, one
human clinical study has demonstrated that exposure to DE particles is capable of inducing de novo
sensitization to an antigen in atopic individuals.
Toxicological Studies: The 2004 PM AQCD reported numerous studies which provided evidence
for an association of episodic exposure to PM and exacerbation of allergic asthma. The vast majority of
studies conducted since the last review focus on PM2 5. Some of these new studies demonstrate that
existing allergic sensitization confers susceptibility to the effects of PM in rodent models. For example
increased PM retention and enhanced allergic responses were observed in ovalbumin-allergic rats exposed
by inhalation to CAPs (Morishita et al., 2004). Allergic responses were also enhanced in
ovalbumin-allergic mice exposed by inhalation to roadway CAPs (Kleinman et al., 2005; 2007). Greater
responses were observed with exposures closer to the highway. Furthermore, three studies observed AHR
in ovalbumin-allergic mice exposed to DE (Farraj et al., 2006a, b; Matsumoto et al., 2006). Neurotrophins
were found to mediate the response to DE (Farraj et al., 2006b). The 2004 PM AQCD also discussed the
role of PM in acting as an adjuvant to promote allergic sensitization. New studies in mice demonstrated
that inhalation of DE resulted in enhanced allergic sensitization to a common fungus (Liu et al., 2008) and
to ovalbumin (Stevens et al., 2008; Whitekus et al., 2002). In contrast, woodsmoke had minimal effects
on allergic sensitization to ovalbumin in mice (Barrett et al., 2006).
Host Defense
Toxicological studies: The 2004 PM AQCD reported increased susceptibility to infection
following exposure to PM. Inhalation of CAPs prior to bacterial infection was found to increase the
bacterial burden in aged rats (Zelikoff et al., 2003). ROFA, administered by intratracheal instillation, was
also found to diminish bacterial clearance (Antonini et al., 2002). New studies evaluating the effects of
PM on host defense focused on PM2 5. Inhalation exposure of mice to DE resulted in an increased
susceptibility to influenza infection (Ciencewicki et al., 2007) and respiratory syncytial virus (Harrod et
al., 2003). Decreased levels of surfactant proteins known to play an important role in viral clearance were
observed in both of these studies (Ciencewicki et al., 2007; Harrod et al., 2003). In contrast, woodsmoke
December 2008
6-191
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
inhalation had no effect on bacterial clearance in four rodent species, including one which was
compromised (Reed et al., 2006).
Hospital Admissions and ED Visits
Epidemiologic Studies: The majority of new evidence relating to the effect of PM2.5 on HAs for
respiratory causes comes from several recent MCAPS analyses of older adults. MCAPS investigators
report consistent associations of PM25 with COPD and respiratory infections with heterogeneity in these
estimates explained by regional and seasonal differences (Bell et al., 2008a; Dominici et al., 2006). In
another recent study, SOPHIA investigators observed non-significant increases in ED visits for respiratory
causes with PM2.5 among all ages but PM2 5 data were available for fewer years than PMi0 data (Peel et
al., 2005). However, a multicity study in France and several single-city studies that were conducted in
Canada that show weaker associations between hospitalization and acute exposure to PM2 5 compared to
PMio-2 5 (Chen et al., 2004b; Chen et al., 2005b; Fung et al., 2006; Host et al., 2007; Lin et al., 2002b; Lin
et al., 2005; Yang et al., 2004b). All were studies of hospitalization for respiratory diseases, though studies
differed in age group and respiratory endpoint.
Barnett et al. (2005) reported increased HAs with PM2 5 concentration in 7 cities in Australia and
New Zealand among children. Heterogeneous effect estimates were observed across single city studies of
asthma admissions and ED visits (Babin et al., 2007; Barnett et al., 2005; Ito et al., 2007; Peel et al.,
2005; Slaughter et al., 2005).
A report from the SOPHIA study in Atlanta evaluated associations between short-term fine particle
component exposures, using ARIES data, and visits for respiratory diseases (Peel et al., 2005). No
significant associations were reported between any component and respiratory visits, except for an
association between OC and emergency department visits for pneumonia (Peel et al., 2005). Medical
visits for asthma in children and lower respiratory infections (all ages) were associated with the EC and
OC components of fine particles in Atlanta, but no associations were reported with sulfates or acidity
(Sinclair and Tolsma, 2004). Metals were positively associated with medical visits for lower respiratory
infection, but not for other outcomes. For adult asthma and upper respiratory infections, there were no
significant positive associations with any of the fine PM components; however, sulfates were negatively
associated with upper respiratory infection visits (Sinclair and Tolsma, 2004).
In the SOPHIA study in Atlanta, PM2 5 from mobile sources, biomass burning and sulfate-rich
secondary PM2 5 were associated with a 2-4% increase in respiratory hospital visits (Sarnat et al., 2008).
Biomass was associated with total respiratory hospitalizations and vehicle emissions with childhood
asthma hospitalizations (Andersen et al., 2007b). PM from traffic was linked to pneumonia HAs in
Boston (Zanobetti and Schwartz, 2006). Vegetative burning was associated with respiratory HAs in
December 2008
6-192
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Spokane despite the lack of association with PM size fractions studied (Schreuder et al., 2006; Slaughter
et al., 2005).
6.3.9.4. Ultrafine Particles
Causal Determination
Although a very limited number of epidemiologic studies have provided some evidence of an
association between short-term exposure to ultrafine particles and respiratory symptoms as well as asthma
hospitalizations, these findings have been inconsistent across studies. The effect of controlled exposures
to ultrafine particles has not been extensively examined in humans. Two human clinical studies have
observed small ultrafine particle-induced decreases in pulmonary function, however, no increases in
respiratory symptoms or pulmonary inflammation have been reported. The results from animal
toxicological studies examining the effect of ultrafine particles on respiratory morbidity are mixed, and
the interpretation of the findings limited by a relative lack of data. Thus, current collective evidence is
inadequate to determine if a causal relationship exists between relevant UFP exposure and
short-term respiratory morbidity.
Respiratory Symptoms and Medication Use
Epidemiologic Studies: One study found positive associations with UFP and wheeze and
medication use (von Klot et al., 2002), though another found no association with any respiratory
symptoms (de Hartog et al., 2003).
Human Clinical Studies: The 2004 PM AQCD did not include any human clinical studies that
assessed respiratory symptoms following controlled exposure to ultrafine PM. Three new studies have
found no increase in respiratory symptoms following exposure to ultrafine zinc oxide (Beckett et al.,
2005), CAPs (Gong et al., 2008) or EC (Pietropaoli et al., 2004).
Pulmonary Function
Human Clinical Studies: There were no human clinical studies presented in the 2004 PM AQCD
that evaluated the effect of ultrafine PM on pulmonary function. Two new studies (Gong et al., 2008;
Pietropaoli et al., 2004) have demonstrated ultrafine-induced decrements in pulmonary function,
measured as decreases in maximal mid-expiratory flow and CO diffusing capacity (EC) and decreases in
arterial oxygen saturation and FEV, (CAPs). One additional ultrafine CAPs study found no effect of
exposure on FVC, FEV,, or CO diffusing capacity (Samet et al., 2007).
December 2008
6-193
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Toxicological Studies: Very few studies involving ultrafine PM were reviewed in the 2004 PM
AQCD. None of these evaluated pulmonary function effects. Since the last review, one study reported no
changes in AHR following a multiday exposure of mice to ultrafine iron-soot PM (Last et al., 2004).
Pulmonary Inflammation
Human Clinical Studies: No human clinical studies evaluating the relationship between exposure
to ultrafine particles and pulmonary inflammation were presented in the 2004 PM AQCD. Three new
studies have not observed increases in any markers of pulmonary inflammation following exposure to
ultrafine CAPs, zinc oxide, or EC.
Toxicological Studies: The 2004 PM AQCD reported pulmonary inflammation in healthy rats and
mice exposed by inhalation to ultrafine and fine particles of Ti02 at extremely high concentrations.
Ultrafine particles were more inflammogenic than fine particles. Since then, one study demonstrated an
inflammatory response in healthy rats at a much lower concentration of carbon black (1400 (ig/m3,
(Gilmour et al., 2004c), with ultrafine particles more inflammogenic than fine particles. Ultrafine carbon
black particles, at a concentration an order of magnitude lower, was found not to result in pulmonary
inflammation in 2 animal models of aged, compromised rats (Elder et al., 2004b). Furthermore, no
pulmonary inflammation was observed in healthy rats exposed to ultrafine iron-soot particles (Last et al.,
2004). However, a single day exposure to ultrafine CAPs resulted in pulmonary inflammation in the rat
monocrotaline model of pulmonary hypertension (Lei et al., 2004b).
Pulmonary Injury
Toxicological Studies: No histophathological responses were seen in adult mice exposed to
ultrafine iron-soot particles. In contrast, exposure of neonatal rats to ultrafine iron-soot particles during
the neonatal period resulted in a significantly reduced rate of cell proliferation in the proximal alveolar
region (Pinkerton et al., 2004). Alveolar septation and growth were unaffected in this study. The authors
suggest the possibility of greater susceptibility to air pollution during the critical period of postnatal lung
development.
Other than the above study, the relationship between exposure to ultrafine PM and markers of
pulmonary injury has not been evaluated in inhalation studies. However an important series of studies has
evaluated the relative toxicity of PM size fractions by measuring markers of injury, as well as markers of
inflammation, in BALF of mice following intratracheal instillation or aspiration of ambient PM from a
variety of European and U.S. cities. Ultrafine PM was less injurious than the other size fractions. Gilmour
et al. (2004a) evaluated the relative toxicity of Montana coal fly ash of differing size fractions. In contrast
December 2008
6-194
DRAFT-DO NOT QUOTE OR CITE
-------
1 to the results above, the ultrafine fraction fractions from the Montana coal induced greater injury and
2 inflammation than the PM10_2.5 fraction.
Allergic Responses
3 Toxicological Studies: Ultrafine CAPs were more potent than fine CAPS in exacerbating allergic
4 responses in one study (Kleinman et al., 2005). Exposure to ultrafine iron-soot particles resulted in an
5 increased number of goblet cells in allergic mice (Last et al., 2004).
Hospital Admissions and ED Visits
6 Epidemiologic Studies: Studies conducted in Copenhagen, Denmark reported associations with
7 ultrafine particles. Both accumulation mode and number concentration (< lOOnm) were associated with
8 childhood asthma admissions (Andersen et al., 2007b). Associations with ultrafine were not observed by
9 SOPHIA investigators (Peel et al., 2005).
6.4. Central Nervous System Effects
10 While evidence of an effect of PM on the central nervous system was not presented in the 2004 PM
11 AQCD, a limited number of recent human clinical and toxicological studies have provided some support
12 to suggest that exposure to PM may be associated with changes in neurological function. Two
13 epidemiologic studies evaluated the effect of ambient PM on the central nervous system (Calderon-
14 Garciduenas et al., 2008; Suglia et al., 2008). These studies examined long-term exposure to non-specific
15 PM indicators and are detailed in Annex E.
6.4.1. Human Clinical Studies
16 In a recent controlled human exposure study, Cruts et al. (2008) exposed 10 healthy males (18-39
17 years old) to filtered air and dilute DE (300 |ig/m3 particulate concentration) for 1-h using a randomized
18 crossover study design. Changes in brain activity were measured during and following exposure using
19 quantitative electroencephalography (QEEG). Exposure to DE was observed to significantly increase the
20 median power frequency (MPF) in the frontal cortex during exposure, as well as in the hour following the
21 completion of the exposure. While this study does provide some evidence of an acute cortical stress
22 response to DE, it is important to note that the QEEG findings are very nonspecific, and could have been
23 caused by factors other than diesel PM such DE gases (e.g., CO, NO and N02) or the odor of the DE.
December 2008
6-195
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
6.4.2. Toxicological Studies
Evidence is mounting that the nervous system may be a critical target of PM and that adverse
health effects may result from PM exposure. Whether these health effects are a direct or indirect effect of
PM has not yet been established. One hypothesis suggests that ultrafine PM which deposits onto nasal
olfactory epithelium enters the central nervous system by axonal olfactory transport to the olfactory bulb
and leads to a cascade of effects involving inflammatory cytokines and ROS. An increased potential for
neurodegenerative processes may ensue. Evidence for translocation of ultrafine PM to the olfactory bulb
via olfactory neurons is discussed in Chapter 4, but its relevance to CNS health effects is unknown.
Another hypothesis suggests that brain inflammation occurs secondarily to PM-mediated systemic
inflammation. Finally, it has been suggested that PM-stimulation of autonomic nervous system receptors
in the respiratory tract results in inflammatory or other effects in the central nervous system. This is an
emerging field with many unknowns.
Calderon-Garciduenas et al. (2003) conducted a long-term exposure observational study in mongrel
dogs from Mexico City and Tlaxcala. DNA damage and inflammation in the brain and respiratory tract
were evaluated in dogs living in Mexico City (exposed group) and dogs living in Tlaxcala (control
group). These cities are similar in altitude but differ in air pollutant levels. Measurements of air pollutant
levels were presented only for Mexico City, the more polluted city. Statistically significant greater levels
of apurinic/apyrimidinic sites (an indicator of DNA damage) were observed in the olfactory bulbs and
hippocampus of Mexico City dogs compared with controls. These differences were not seen in other brain
regions examined or in nasal respiratory epithelium. In addition, Mexico City dogs demonstrated greater
histopathological changes in the respiratory and olfactory epithelium of the nasal cavity compared with
controls. Immunohistochemical staining of brain tissue from the Mexico City dogs demonstrated greater
immunoreactivity for NFkB, iNOS, cyclooxygenase-2, glial fibrillatory acidic protein (GFAP), ApoE,
amyloid precursor product and |3-amyloid compared with controls. These results are indicative of
inflammation and stress protein responses. This study has several limitations given that the dogs were of
mixed breeds and of variable ages and that there was no standardization of exposures or diets. However
results indicate a further need for investigating the relationship between air pollution and brain
inflammation.
Several new inhalation studies have provided evidence of CNS effects due to ambient PM
exposures. In one study, Campbell et al. (2005) exposed ovalbumin-sensitized BALB/c mice to filtered air
or near-highway Los Angeles CAPs (a 20-fold concentration of fine+ultrafine or ultrafine only; mean
exposure concentration ultrafine 282.5 (ig/m3 and fine 441.7 (ig/m3 for 4 h/day, 5 days/week for 2 weeks.
The animals were subsequently challenged with ovalbumin to elicit an allergic response in the lungs;
December 2008
6-196
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
brain tissue was obtained 1 day later. Exposure to CAPs, but not filtered air, resulted in activation of the
immune-related transcription factor NF-kB and the cytokines TNFa, and IL-la in brain, demonstrating
proinflammatory responses that could contribute to neurodegenerative disease. While this study
demonstrates CAPs effects in an allergic animal model, futher studies are required to determine whether
these responses also occur in non-allergic animals.
In a second study, control or ovalbumin-sensitized and challenged Brown Norway rats were
exposed for 8-h to filtered air or fine CAPs (Grand Rapids, MI; 500 (ig/m3 fine PM) (Sirivelu et al.,
2006). Brain tissue was obtained 1 day later. CAPs exposure resulted in brain region-specific modulation
of neurotransmitters. In animals which were not pretreated with ovalbumin, statistically significant
increases in norepinephrine were observed in the paraventricular nucleus and olfactory bulb of
CAPs-exposed rats compared with filtered air controls. In animals which were pretreated with ovalbumin,
a statistically significant increase in dopamine was observed in the medial preoptic area in CAPs-exposed
rats compared with controls. Furthermore, exposure to CAPs resulted in a statistically significant increase
in serum corticosterone. These data suggest that the hypothalamo-pituitary-adrenal axis (i.e. stress axis)
may be activated by PM exposure, causing aggravation of allergic airway disease. The authors discuss the
possible role of the olfactory bulb in mediating neuroendocrine control of autonomic activities involved in
respiratory and cardiovascular functions; however further investigations are required to clarify these
relationships.
In a third study, normal (C57BL/6) mice and ApoE"" mice were exposed to Tuxedo, NY, fine CAPs
for 4 mo (March, April or May through September 2003) (Veronesi et al., 2005). The average PM2 5
exposure concentration was 110 |ig/nr\ CAPs exposure resulted in a statistically significant decrease in
domaminergic neurons, measured by tyrosine hydroxylase immunoreactivity, in the substantia nigra of
ApoE ""mice but not in control mice. This population of neurons is targeted in neurodegenerative diseases
such as Parkinson's. Furthermore, a statistically significant increase in GFAP immunoreactivity, a marker
for astrocytes, was observed in the nucleus compacta of CAPs-exposed ApoE" " mice compared to
air-exposed ApoE" " mice. These results suggest that the ApoE" " mice, a genetic model for increased
oxidative stress, are susceptible to PM-induced neurodegeneration. Evidence for brain oxidative stress has
also been found in normal animals following intratracheal instillation of high concentrations of fine CAPs
from Taiyuan, China (Liu and Meng, 2005) and of gasoline exhaust (Che et al., 2007) and following
chronic exposure to ROFAby intranasal instillation (Zanchi et al., 2008).
In summary, PM may produce adverse effects in the CNS by direct or indirect mechanisms which
are at present incompletely understood. Two recent short-term fine CAPs inhalation studies demonstrate
proinflammatory responses in the brain and brain region-specific modulation of neurotransmitters and
suggest the involvement of neuroimmunological pathways. One recent chronic fine CAPs inhalation
December 2008
6-197
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
study demonstrates loss of dopaminergic neurons in the substantia nigra and suggests that oxidative stress
contributes to neurodegeneration. Veronesi et al. (2005) have noted that the brain is very vulnerable to the
oxidative stress induced by PM due to the brain's high energy demands, low levels of endogenous free
radical scavengers, and high content of lipids and proteins. PM-mediated upregulation of inflammatory
cytokines and mediators may also contribute to neurodegeneration. Further investigations are required to
substantiate these mechanisms.
6.4.3. Summary and Causal Determination
A single human clinical study provides some evidence of an acute cortical stress response to diesel,
though these findings are nonspecific and could have been caused by DE gases rather than diesel PM.
Recent animal toxicological studies of CAPs have demonstrated proinflammatory responses in the brain,
brain region-specific modulation of neurotransmitters and loss of dopaminergic neurons in the substantia
nigra; however, the mechanisms underlying these effects need to be substantiated. Though the effect of
ambient air pollution on CNS outcomes has recently begun to draw more attention, the evidence for a
CNS effect associated with PM is limited. While most available studies have evaluated the effects of fine
particle exposures, there is insufficient evidence to draw conclusions regarding effects of specific PM size
fractions or components. Overall, the evidence is inadequate to determine if a causal relationship
exists between relevant exposures to PM10, PM2 5, PM10-2.5, ultrafine particles, or specific PM
components and CNS outcomes.
6.5. Mortality Associated with Short-Term Exposure
The relationship between short-term exposure to PM and mortality is an important issue that has
been extensively addressed in previous PM assessments (U.S. EPA, 1982, 1996, 2004). A positive
association between PM concentration and mortality was consistently demonstrated across studies cited in
the 2004 PM AQCD. Although effect estimates have been shown to be dependent on regional and
seasonal differences, recent studies have provided additional support to previous findings. The current
body of evidence examines the association between short-term exposure to PM of various size fractions
(i.e., PM10, PM10-2 5, PM2 s and ultrafine particles [0.01-0.1 |_im |) and mortality through the use of
time-series and/or case-crossover studies. Both study designs aim to disentangle the PM-mortality effect
through either complex modeling (i.e., time-series) or matching strategies (i.e., case crossover). The
scientific community has also continued to search for the specific PM components and sources that are
December 2008
6-198
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
responsible for the health effects attributed to short-term exposure to PM, along with the
concentration-response relationship that most adequately explains the effect of PM on human health.
6.5.1. Summary of Findings from 2004 PM
The 2004 PM AQCD found strong evidence that ambient coarse (PMi0) and fine (PM2 5), or one or
more PM2 5 components, acting alone and/or in combination with gaseous copollutants, are associated
with total (non-accidental) mortality and various cause-specific mortality outcomes. For PMi0, several
multicity studies in the U.S., Canada, and Europe provided strong support for this conclusion, reporting
associations with total mortality highlighted by effect estimates ranging from -0.2 to 0.7% (per 10 |ig/m3
increase in PMi0) (U.S. EPA, 2004). Numerous studies also reported PMi0 associations with
cause-specific mortality, specifically cardiovascular- and respiratory-related mortality. For PM2 5, the
strength of the evidence varied across endpoints, with relatively stronger evidence for associations with
cardiovascular compared to respiratory endpoints. The resulting effect estimates reported from the U.S.
and Canadian based studies (both multi- and single-city) analyzed for these two endpoints ranged from
1.2 to 2.7% for cardiovascular-related mortality and 0.8 to 2.7% for respiratory-related mortality, per
10 |ig/m3 increase in PM2 5 (U.S. EPA, 2004). In regards to thoracic coarse particles (PM10.2 5), the PM
AQCD found a limited body of evidence that was suggestive of associations between short-term exposure
to ambient PM10_2.5 and various mortality outcomes (e.g., 0.08 to 2.4% increase in total [non-accidental]
mortality per 10 |ig/m3 increase in PMi0.2.5). The positive effect estimates obtained from studies that
analyzed the association between PMi0.2 5 and mortality resulted in the conclusion that PMi0.2 5, or some
constituent component(s) (including those on the surface) of PMi0.2.5, may contribute, in certain
circumstances, to increased human health risks.
Some additional studies examined the association between specific PM2 5 chemical components and
mortality. These studies observed associations for sulfate, nitrate, and CoH, but not crustal particles. The
strength of the association for each component varied from city to city (U.S. EPA, 2004). Source-oriented
analyses were also conducted to identify specific source-types associated with mortality. These studies
implicate fine particles from anthropogenic origin, such as motor vehicle emissions, coal combustion, oil
burning, and vegetative burning, as being important in contributing to increased mortality (U.S. EPA,
2004).
December 2008
6-199
DRAFT-DO NOT QUOTE OR CITE
-------
Table 6-13. Overview of U.S. and Canadian multicity PM studies analyzed in the 2004 PM AQCD
and the PM ISA"
Reference
Location
Mean Concentration (|jg/m3)c
Upper Percentile:
Concentrations (|jg/m3)
PMw
Dominici et al. (2003a)a
90 U.S. cities
15.3-53.2
NR
Burnett and Goldberg (2003)a
8 Canadian cities
25.9
95th: 54
Maximum: 121
Peng et al. (2005)
100 U.S. cities
13-49
50th
75th
27.1
32.0
Dominici et al. (2007b)
100 U.S. cities
13-49
50th
75th
27.1
32.0
Welty and Zeger (2005)
100 U.S. cities
13-49
50th
75th
27.1
32.0
Burnett et al. (2004)
12 Canadian cities
NR
NR
Schwartz (2004b)
14 U.S. cities
23-361
75th: 31 -57
Schwartz (2004c)
14 U.S. cities
23-361
75th: 31 -57
Zeka et al. (2005)
20 U.S. cities
15.9-37.5
NR
Zeka et al. (2006a)
20 U.S. cities
15.9-37.5
NR
PMis
Burnett and Goldberg (2003)=
8 Canadian cities
13.3
95th: 32
Maximum: 86
Dominici et al. (2007b)
100 U.S. cities
NR
NR
Franklin et al. (2007)
27 U.S. cities
9.3-28.5
NR
Franklin et al. (2008)
25 U.S. cities
Winter: 9.6 - 34.4
Spring: 6.7 - 27.6
Summer: 7.6-26.0
Fall: 9.5-32.1
NR
Ostro et al. (2006)
9 California counties
14-29
NR
Burnett et al. (2004)
12 Canadian cities
12.8
NR
PMlO-25
Burnett and Goldberg (2003)=
8 Canadian cities
12.6
95th: 30
Maximum: 99
Burnett et al. (2004)
12 Canadian cities
11.4
NR
Villeneuve et al. (2003)
Vancouver, Canada
6.1
90th: 13.0
Maximum: 72.0
Klemm et al. (2004)
Atlanta, Georgia
9.7
50th: 9.34
75th: 11.94
Maximum: 25.17
Slaughter et al. (2005)
Spokane, Washington
NR
NR
Wilson et al. (2007b)
Phoenix, Arizona
NR
NR
December 2008
6-200
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Reference
Location
Mean Concentration (|jg/m3)c
Upper Percentile:
Concentrations (jjg/m3)
Kettunen et al. (2007)
Helsinki, Finland
Cold season: 6.7d
Warm season: 8.4d
Cold season
50th: 6.7
75th: 12.5
Maximum: 101.4
Warm season
50th: 8.4
75th: 11.8
Maximum: 42.0
Perez et al. (2008)
Barcelona, Spain
Saharan Dust Days: 16.4
Non-Saharan Dust Days: 14.9
Saharan Dust Days
50th: 14.8
75th: 21.8
Maximum: 36.7
Non-Saharan Dust Days
50th: 12.6
75th: 18.9
Maximum: 93.1
a Multicity studies examined in the 2004 PM AQCD
b Because only one multicity study was identified that examined PM - ¦ single-city and International studies that examined PM10-25 were analyzed in this ISA and are included in this
table.
cThe majority of multicity studies examined in the PM ISA provide the mean PM concentration of each individual city, not an overall PM concentration across all cities. As a result, the
range of PM concentrations for a particular study are presented, which represents the lowest and highest mean PM concentrations reported across cities, if an overall mean is not
provided within the study.
d Median PM concentration.
6.5.2. Associations of Mortality and Short-Term Exposure to PM
The recent literature examines the association between short-term exposure to various PM size
fractions (i.e., PMi0, PM10-2.5, PM2 5, ultrafine particles, or species [e.g., OC, EC, transition metals, etc.])
and mortality. This ISA, similar to previous AQCDs, focuses more heavily on multicity studies, and
specifically those conducted in the U.S. and Canada (see Table 6-13). By using this approach it is possible
to: (1) obtain a more representative sample of or insight to the PM-mortality relationship observed across
the U.S.; (2) analyze the association between short-term exposure to PM and mortality at ambient
conditions at or similar to those observed in the U.S.; (3) examine the potential heterogeneity in effect
estimates between cities and regions; and (4) analyze the confounders and/or effect modifiers that may
explain the PM-mortality relationship in the U.S. The one caveat to using this approach for the current
document is that fewer multicity studies have been conducted since the 2004 PM AQCD. This is because
most of the multicity PM studies that were reviewed in the 2004 PM AQCD utilized the multiple-cause of
death files from the National Center for Health Statistics (NCHS). However, due to a change in policy, the
NCHS no longer provides the day of death in national data sets of daily mortality records starting with
data for 2001. As a result, this led to fewer multicity studies analyzing mortality data beyond year 2000,
except for a few studies that requested data directly from state or city agencies. Although this section
focuses on mortality outcomes in response to short-term exposure to PM, it does not evaluate studies that
December 2008
6-201
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
examine the association between PM and infant mortality. These studies are evaluated in Section 7.5,
Reproductive, Developmental, Prenatal and Neonatal Outcomes, although it is possible that short- and
long-term in utero exposures may contribute to infant mortality. In addition, the exposure windows of
interest for this unique health outcome can be difficult to characterize and may span both short- and long-
term periods.
6.5.2.1. PM10
The majority of studies that examined the association between short-term exposure to PM and
mortality focused on effects attributed to PMi0. These studies analyzed the PMio-mortality relationship
through either a time-series or case-crossover design.1
Time-Series Analyses
Mortality associated with short-term exposure to PMi0 has been examined in several updated
time-series analyses of the National Morbidity and Mortality Air Pollution Study (NMMAPS). In the
previous NMMAPS analysis (Dominici et al., 2003a; Samet et al., 2000) of the 1987-1994 data, which
was reviewed in the 2004 PM AQCD, the strongest association was found for non-accidental mortality for
1-day lag, with a combined estimate across 90 cities of 0.21% (95% posterior interval [PI]: 0.09, 0.33) per
10 |_ig/m3 increase in PMi0. The association was found to be robust to the inclusion of other gaseous
copollutants in the regression models, but the investigators found heterogeneity across regions, with the
strongest associations in northeastern cities. In the new updated analyses, the investigators examined
additional issues including: (1) seasonal effect modification; (2) change in risk estimates overtime; and
(3) sensitivity of results to alternative weather models. In addition, a few international multicity studies
were conducted that provide information, which further clarifies the association between PMi0 and
mortality. There has also been an analysis which examines the PM10 concentration-response relationship
in 20 of the NMMAPS cities (see Section 6.5.2.7).
Seasonal Analyses of PMw-mortality associations in 100 U.S. Cities (NMMAPS)
Peng et al. (2005) analyzed the updated NMMAPS data, which consisted of 100 U.S. cities for the
period 1987-2000. In their first stage regression model, for each city, the PM10 effect was modeled to
1 Schwartz (2004b) used a case-crossover study design, but also conducted a time-series analysis to validate the results obtained using the
case-crossover approach.
December 2008
6-202
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
have a sinusoidal shape that completes a cycle in a year but was constrained to be periodic across years
using sine/cosine terms. The authors also considered a model that consisted of PM10-season interactions
using season indicators. Both of these models also included covariates that were used in their earlier
NMMAPS analyses. In the second stage model, the seasonal patterns of PMi0 mortality coefficients were
estimated for seven geographic regions and on average for the entire U.S. Peng et al. (2005) found for
1-day lag, at the national level, season specific increases in non-accidental mortality per 10 (ig/m3
increase in PM10 of: 0.15% (Posterior Interval (PI): -0.08, 0.39), 0.14% (PI: -0.14, 0.42), 0.36% (PI: 0.11,
0.61), and 0.14% (PI: -0.06, 0.34) for winter, spring, summer, and fall, respectively. The corresponding
all-season estimate was 0.19% (PI: 0.10, 0.28). After the inclusion of S02, 03, or N02 in the model with
PMio in a subset of cities (i.e., 45 cities) for which data existed for all pollutants resulted in fairly robust
PMio risk estimates. An analysis by geographic region found a strong seasonal pattern in the Northeast.
Figure 6-14 presents the estimated seasonal pattern of PMi0 risk estimates by region from Peng et al.
(2005), which includes a sensitivity analysis aimed to determine the appropriate number of degrees of
freedom for temporal adjustment. It is clear from Figure 6-14 that the Northeast has the strongest
association, which peaks in the summer and is robust to the extent of temporal adjustment. The industrial
Midwest also shows the summer peak, but with smaller risk estimates. Other regions have either no
seasonal pattern (Southeast) or a suggestion of a spring peak that appears to be sensitive to the extent of
temporal adjustment. On a nationwide basis, the PM10 risk estimates appear to peak between spring and
summer. Overall, this study identified an effect modifier that may be useful in identifying specific
chemical component(s) that are related to specific regions and times of the year.
Change in PMio-mortality associations in NMMAPS data, 1987 to 2000
Dominici et al. (2007b) conducted an analysis of the extended NMMAPS data set with the main
objective of examining any change in short-term PM10-mortality risk estimates during the course of the
study period. They estimated the average PMi0 mortality risk coefficient for 1-day lag using essentially
the same model specification as in their 2003 analysis, separately for three time periods: 1987-1994,
1995-2000, and 1987-2000, for the "eastern U.S." (62 counties), the "western U.S." (38 counties), and
all 100 U.S. counties. To produce national and regional estimates, two-stage hierarchical models were
used as in the previous NMMAPS studies. As shown in Table 6-14, the authors found a continuation of
the PMio-mortality association in the nationwide data for the entire study period. A comparison of the
relative risk estimates for 1987-1994 vs. 1995-2000 suggests weak evidence (not a significant difference)
that short-term effects declined. Most of the decline in the national estimate appears to be attributable to
the eastern U.S. counties. However, the decline in risk estimate in all-cause mortality in the eastern U.S.
appears to be disproportionately influenced by the reduction in risk estimate for the "other" mortality
December 2008
6-203
DRAFT-DO NOT QUOTE OR CITE
-------
1 category (i.e., all-cause minus cardiorespiratory category, which may be 40 to 50% of all-cause deaths in
2 U.S. cities) in the eastern U.S. Likewise, the apparent increase in the risk estimate for all-cause mortality
3 in the western U.S. appears to be affected by the increase in risk estimate for the "other" mortality
4 category. Because it is not clear what specific cause(s) in the "other" mortality category are affected by
5 PM, interpreting the reduction in risk estimates for all-cause mortality requires caution. In contrast, the
6 apparent reductions, -23%, in PMi0 risk estimates for cardio-respiratory deaths were more comparable
7 between the two regions.
cp
Industrial Midwest
Northeast
c 1-5
» 1.0
a>
| 0.5
o
e oo
-V-
South wo st
* 16
® 1.0
100 200 300
100 200 300
100 20© 300
10© 200 300
Day in year
Source: Peng et al. (2005)
Figure 6-14. National and regional estimates of smooth seasonal effects for PM10 at a 1-day lag and
their sensitivity to the degrees of freedom assigned to the smooth function of time in
the updated NMMAPS data 1987-2000. Note: The degrees of freedom chosen were 3
df (short-dashed line), 5 df (dotted line), 7 df (solid line), 9 df (dotted-and-dashed line),
and 11 df (long-dashed line) per year of data.
8 In addition, the investigators estimated time-varying PMi0 mortality risk as a linear function of
9 calendar time for the period 1987-2000, producing the percentage rate change in PMi0 risk estimate with
10 a change in time of 1 year. The estimated rate of decline in slope for all-cause and the combination of
11 cardiovascular and respiratory mortality were -0.012 (PI: -0.037, 0.014) and -0.016 (PI: -0.058, 0.027),
December 2008
6-204
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
respectively. The authors also estimated a PM2 5 mortality risk for the period 1999-2000 (discussed in
Section 6.5.2.2.).
Table 6-14. NMMAPS national and regional percentage increase in all-cause, cardio-respiratory,
and other-cause mortality associated with a 10 pg/m3 increase in PM10 at lag 1 day for
the periods 1987-1994,1995-2000, and 1987-2000.
1987-1994 95% PI 1996-2000 95% PI 1987-2000 95% PI
ALL CAUSE
East 0.29 0.12,0.46 0.13 -0.19,0.44 0.25 0.11,0.39
West 0.12 -0.07,0.30 0.18 -0.07,0.44 0.12 -0.02,026
National 0.21 0.10,0.32 0.18 0.00,0.35 0.19 0.10,0.28
CARDIORESPIRATORY
East 0.39 0.16,0.63 0.30 -0.13,0.73 0.34 0.15,0.54
West 0.17 -0.07,0.40 0.13 -0.23,0.50 0.14 -0.05,0.33
National 0.28 0.14,0.43 0.21 -0.03,0.44 0.24 0.13,0.36
OTHER
East 0.21 -0.03,0.44 0.00 -0.49,0.50 0.15 -0.09,0.39
West 0.09 -0.21,0.38 0.23 -0.15,0.62 0.11 -0.10,0.33
National 0.15 -0.02,0.32 0.17 -0.07,0.41 0.15 0.00,0.29
Source: Dominici et al. (2007b)
The objective of the Dominici et al. (2007b) study described above was motivated by
accountability research, the idea of measuring the impact of policy interventions. However, unlike the
intervention studies conducted in Hong Kong (Hedley et al., 2002) and Dublin, Ireland (Clancy et al.,
2002) that were reviewed in the 2004 PM AQCD, this study was not designed to estimate a reduction in
mortality in response to a sudden change in air pollution. In fact, the figure of observed trend in PM10
levels presented in the Dominici et al. (2007b) study indicates that the decline in PMi0 levels during the
study period was very gradual, with much of the decline appearing in the first few years (median values
of ~33 (ig/m3 in 1987 to ~25 (ig/m3 in 1992, then down to ~23 (ig/m3 in 2000). A flaw in the use of the
time-series study design for this type of analysis is that it adjusts for long-term trends, and, therefore, does
not estimate the change in mortality in response to the gradual change in PMi0. The apparent change,
though weak, in the PMi0 risk estimates may also reflect a potential change in the chemical composition
of PMio. The study listed a number of PMio-related air pollution control programs that were implemented
between 1987 and 2000. Some of these programs, such as the Acid Rain Control Program, did result in
major reductions in emissions, and, therefore, could have contributed to the results observed, but the
analytic approached used in the study does not allow for a systematic analysis of the effect of air pollution
policies on the risk of mortality.
December 2008
6-205
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Sensitivity of NMMAPS PM10 risk estimates to alternative weather models
Welty and Zeger (2005) analyzed the updated NMMAPS 100 U.S. cities data to examine the
sensitivity of PM10 mortality risk estimates to alternative weather models that consider longer lags. All of
the previous NMMAPS analyses only considered temperature and dew point up to 3-day lags. In this
analysis, the authors considered various forms of a constrained distributed lag model: (1) containing a
step function of temperature with steps at lag 0, 2, 7 and extended to 14 days; (2) similar to (1) but with
time-varying coefficients to change over season and study period; and, (3) containing a smooth function
to account for non-linearity in the temperature-mortality relationship. With a combination of degrees of
freedom for temporal trends and the number of distributed lags, more than 20 models were applied to
each of the three lag days (0, 1, and 2) of PMi0. These city-specific risk estimates were then combined
across the 100 cities in the second stage Bayesian model. The combined PMi0 risk estimates were
generally consistent within the lag. In particular, the risk estimates for non-accidental mortality for lag
1-day ranged between 0.15% and 0.25% per 10 |_ig/m3 increase in PMi0, and were always statistically
significant regardless of the model used. In addition, the range of these point estimates across the models
was found to be much narrower than the regression posterior intervals. Thus, the PMi0 risk estimates at
lag 1 day were robust to alternative temperature models that considered temperature effects lasting up to a
two-week period.
In summary, the above three analyses of the updated NMMAPS data provided useful information
on PM10 mortality risks, resulting in the following conclusions: (1) estimated PM10 risk is particularly
high in the northeast and in the summer; (2) there remains an overall PM10-mortality association in the
1987-2000 time period as well as the 1995-2000 time period; (3) there is a weak indication that PM10
mortality risk estimates are declining; and (4) PM10 risk estimates were not sensitive to alternative
temperature models.
PM10 Mortality Studies Conducted in Canada and Europe
Burnett et al. (2004) examined the association between mortality and various air pollutants in 12
Canadian cities, and reported that the most consistent association was found for N02. For this analysis,
PM was measured every 6th day for the majority of the study period, and the PMi0 concentrations used in
the study represent the sum of the PM2 5 and PMi0_2 .5 measurements obtained. The authors found that the
simultaneous inclusion of N02 and PMi0 in a model, on those days with PM data, greatly reduced the
PM10 association with non-accidental mortality, from 0.47% (95% CI: 0.04-0.89) to 0.07% (95% CI:
-0.44 to 0.58) per 10 (ig/m3 increase. The previous Canadian multicity analysis (Burnett and Goldberg,
2003), a re-analysis of Burnett et al. (2000) reviewed in the 2004 PM AQCD, did not consider gaseous
December 2008
6-206
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
pollutants. Thus, PM10 risk estimates in the Canadian data appear to be more sensitive to N02 than those
estimates reported in U.S. studies.
The association between PM10 and mortality in Europe was also reviewed in the 2004 PM AQCD
through Katsouyanni et al. (2003), which presented results from the Air Pollution and Health: a European
Approach (APHEA2) study, a multicity study that examined PMi0 effects on total mortality in 29
European cities. Analitis et al. (2006) published a brief report on effect estimates for cardiovascular and
respiratory deaths also based on the 29 European cities, within the APHEA2 study. They reported for the
average of 0- and 1-day lags, PMi0 risk estimates per 10 (ig/m3 of 0.76% (95% CI: 0.47-1.05) for
cardiovascular deaths and 0.71% (95% CI: 0.22-1.20) for respiratory deaths in random effects models.
Case-Crossover Analyses
Since the 2004 PM AQCD investigators have used the case-crossover study design more frequently
as an alternative to time-series analyses to examine the association between short-term exposure to PM
and mortality. This study design allows for the control of seasonal variation, time trends, and slow time
varying confounders without the use of complex models. However, similar to any study design, biases can
be introduced into the study depending on the control (i.e., referent) period selected (Janes et al., 2005).
The multicity case-crossover analyses discussed below match cases (i.e., days in which a death occurred)
to controls (i.e., days in which a death did not occur), to control for (1) seasonal patterns and gaseous
pollutants, or (2) temperature. In addition the studies attempt to examine the heterogeneity of effect
estimates through the analysis of individual-level and city-specific effect modification.
PMw and Mortality in 14 U.S. Cities: Controlling for Temperature
Schwartz (2004b) investigated the PMio-mortality association in 14 U.S. cities for the years
1986-1993 (some cities started in later years because of PM10 data availability) using a case-crossover
study design. Note that in this analysis, four more cities (Boulder, CO; Cincinnati, OH; Columbus, OH;
and Provo-Orem, UT) were added to the cities Schwartz (2003) previously analyzed using a time-series
study design. These cities were chosen for this analysis because they collected daily PM10 data, unlike
most U.S. cities, which only monitor PMi0 every six days. Lag 1-day PMi0 risk estimates were computed
using several methods. Models (1) (i.e., the main model) and (2) were constructed from a case-crossover
analysis with bidirectional control days (7-15 days before and after the case). Model (1) obtained
city-specific estimates in the first stage analysis, followed by a second stage random-effects model to
obtain a combined estimate. Model (2) is the same as model (1), but consisted of a single stage model,
which included data from all 14 cities. Models (3) and (4) were also constructed from a case-crossover
analysis, but used time-stratified control days (i.e., matched on season and temperature within the same
December 2008
6-207
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
degree in Celsius). Model (3) obtained single-city estimates in the first stage analysis, followed by a
second stage random-effects model to obtain combined estimates. Model (4) used the same approach as
model (3), but consisted of a single stage model including data from all 14 cities. The final model, (5),
consisted of a two-stage Poisson time-series model, which produced city-specific estimates in the first
stage, and combined estimates across cities in the second stage. In the main model, (1) above, the
estimated excess risk for non-accidental mortality was 0.36% (CI: 0.22, 0.50) per 10 |_ig/m3 increase in
PMio. The other models yielded a similar magnitude of effect estimates, ranging from 0.32% (model 2) to
0.53% (model 4). Thus, the methods used to select control days and adjust for weather in the
case-crossover design did not result in major differences in effect estimates, and in addition, were
comparable to the estimates obtained from the time-series analysis, 0.40% (model 5).
PMw and Mortality in 14 U.S. Cities: Controlling for Gaseous Pollutants
In a subsequent analysis, Schwartz (2004a) analyzed the same 14 cities data described above, using
a case-crossover design, to investigate the potential confounding effects of gaseous pollutants. For each
case day, control days were selected from all other days of the same month of the same year. In addition,
case days were matched to control days that had gaseous pollutant levels that were within a defined
concentration: 1 ppb, 1 ppb, 2 ppb, and 0.03 ppm for S02, N02, 1-h max 03, and CO, respectively. Unlike
the study described above (Schwartz, 2004b), in this analysis, the excess risk was estimated for the
average of 0- and 1-day lag PMi0 (rather than 1-day lag). In addition, apparent temperature (a composite
index of temperature and humidity) was used rather than temperature and humidity individually. The
case-crossover analysis was conducted in each city, and a combined estimate was computed in a
second-stage random effects model. The number of cities analyzed varied across pollutants depending on
the availability of monitors. The study reported PM10 risk estimates for non-accidental mortality of 0.81%
(CI: 0.47, 1.15), 0.78% (CI: 0.42, 1.15), 0.45% (CI: 0.12, 0.78), and 0.53% (CI: 0.04, 1.02) per 10 (ig/m3
increase, for the analysis matched by S02 (10 cities), N02 (8 cities), 03 (13 cities), and CO (13 cities),
respectively.
Schwartz (2004c) only presented PMi0 risk estimates matched by gaseous pollutants, therefore, it is
unclear in this analysis how matching by gaseous pollutants affected (i.e., reduced or increased)
unmatched PMi0 risk estimates. The estimates reported were computed using the average of 0- and 1-day
lagged PMio and, therefore, cannot be directly compared to the 1-day lag PMi0 risk estimates obtained in
the Schwartz (2004b) 14-city study described above. The estimates reported in Schwartz (2004c) are
generally larger than those obtained in the Schwartz (2004b) analysis, which was expected since the
Schwartz (2004c) analysis used two-day average PMi0. However, the estimates reported in Schwartz
(2004c) are comparable to the average of 0- and 1-day lagged PMi0 risk estimate for non-accidental
December 2008
6-208
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
mortality (0.55% [CI: 0.39, 0.70]) per 10 (ig/m3 increase from the 10-city study (Schwartz, 2003), which
was reviewed in the 2004 PM AQCD. Overall, Schwartz (2004c) provided an alternative method to assess
the influence of gaseous copollutants. The results suggest that PM10 is significantly associated with
all-cause mortality after controlling for each of the gaseous copollutants.
PMw and Mortality in 20 U.S. Cities: City-level Effect Modification
Zeka et al. (2006a) expanded the 14 cities analysis conducted by Schwartz (2004b, c) to 20 cities,
added more years of data (1989-2000), and investigated PMi0 effects on total and cause-specific mortality
using a case-crossover design. Individual 0-, 1-, and 2-day lags as well as an unconstrained distributed lag
model with 0, 1, and 2 lag days were examined. For each case day, control days were defined as every
third day in the same month of the same year, to eliminate serial correlation. The authors also investigated
potential effect modifiers in the second stage regression using city-specific variables including percent
using air conditioning, population density, standardized mortality rates, the proportion of elderly in each
city, daily minimum apparent temperature in summer, daily maximum apparent temperature in winter, and
the estimated percentage of primary PMi0 from traffic sources.
The investigators found that, for all-cause (non-accidental) mortality, lag 1-day showed the largest
risk estimate (0.35% [CI: 0.21, 0.49] per 10 (.ig/nr1) among the individual lags. Respiratory mortality
exhibited associations at lag 0, 1, and 2 days (0.34, 0.52, and 0.51%, respectively), whereas
cardiovascular mortality was most strongly associated with PMi0 at lag day 2 (0.37%). The sum of the
distributed lag risk estimates (e.g., 0.45% [CI: 0 .25, 0.65] for all-cause mortality) was generally larger
than those for single-day lag estimates. The excess risk estimates for single-day lags for specific
respiratory and cardiovascular causes had generally wider confidence intervals due to their smaller daily
mortality counts, but some of the categories showed markedly larger estimates when included in the
combined distributed lag model (e.g., pneumonia 1.24% [CI: 0.46, 2.02]). As shown in Figure 6-15, Zeka
et al. (2005) also found evidence indicative of several PMi0 effect modifiers including higher population
density and the estimated percentage of primary PMi0 from traffic. When 25th vs. 75th percentiles of
these city-specific variables were evaluated, the estimated percent increase in mortality attributed to PMi0
appears to contrast substantially (e.g., 0.09 vs. 0.52% for variance of summer time apparent temperature).
December 2008
6-209
DRAFT-DO NOT QUOTE OR CITE
-------
oj^_
J-
o
o
tr,
O
£=
1
0.8
0.6
0.4
0.2
0
-0.2
-0.4
1000 3000
2
(count/mile )
25% 75%
25% 75%
25% 75%
Popu lation
density
Variance of
summer AT
Mean of
winter AT
% Primary
PMl 0 from
traffic
City effect modifiers
Source: Zeka et al. (2005)
Figure 6-15. Effect modification by city characteristics in 20 U.S. cities. Note: The two estimates
and their CI for each of the modifying factors represent the percentage increase in
mortality for a 10 pg/m3 increase in PM10, for the 25th percentile, and 75th percentile of
the modifier distribution across the 20 cities.
1 The effect modifiers investigated by Zeka et al. (2005) consisted of city-specific variables. Some of
2 these variables are ecological in nature, and therefore, interpreting the meaning of "effect modification"
3 requires some caution. As the investigators pointed out, the population density and the estimated
4 percentage of primary PMi0 from traffic were correlated in this data set (r = 0.65)1. These variables may
5 also be a surrogate for another or composite aspects of "urban" characteristics. Thus, the apparent effect
6 modification by traffic associated PMi0 needs further investigation. Interestingly, the percent of homes
7 with central air conditioning was not a significant effect modifier of PMi0 risk estimates, which questions
1 The correlation coefficient was calculated based on the numbers provided in Table 1 of Zeka et al. (2005).
December 2008 6-210 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
the impact of reduced ventilation rates on PM exposure. Overall, this study presented PM10 risk estimates
that are consistent with those found in other analyses, but also provided new information on the risk
estimated for broad and specific respiratory and cardiovascular mortality, along with possible effect
modifying city characteristics.
PMw and Mortality in 20 U.S. Cities: individual-level Effect Modification
Zeka et al. (2006a) examined individual-level, instead of city-specific, effect modification of
PMio-mortality associations in the 20 U.S. cities described above using the same case-crossover design.
City-specific estimates were obtained in the first stage model, followed by a second stage model which
estimated the overall effects across all cities. Figure 6-16 shows PMi0 excess risks by four of the
individual characteristics examined in the study (i.e., gender, race, age group, and education). It should be
noted that the lag and averaging of days of associations reported varied across the outcomes: all-cause and
heart disease deaths used the average of lag 1 and 2 days; respiratory deaths used the average of lag 0
through 2 days; myocardial infarction deaths used lag 0 day; and stroke deaths used lag 1 day. PMi0 risk
estimates do not appear to differ by gender or by race. However, significant differences were found for the
youngest vs. oldest age groups for all-cause and heart disease mortality. For all-cause mortality, the level
of education appeared to be inversely related to the PMi0 risk estimates, but this observation was not
statistically significant. The study also examined effect modification by location of death
("out-of-hospital" vs. "in-hospital") and season (see Figure 6-17). The "out-of-hospital" deaths showed
larger PM10 risk estimates than were found for "in-hospital deaths" with a significant contrast per
10 (ig/m3 for all-cause (0.71% vs. 0.22%) and heart disease (0.93% vs. 0.15%) deaths. Stroke deaths also
showed a significant contrast (0.87% vs. 0.06%, not shown in Figure 6-17).
December 2008
6-211
DRAFT-DO NOT QUOTE OR CITE
-------
Gender
Race
cc
ifi
i/>
(D
O
3
JJ
in
o
• Male
° Female
£
o
g)
n
0
1
DC
ifl
l/l
;
_U
cc
Education
LTj
CD
q
O
in
~
it.
Medium
CL
Ifi
CD
GC
Figure 6-16. PM10 risk estimates (per 10 pg/m3) by individual-level characteristics. The risk
estimates and 95% confidence intervals were plotted using numerical results from
tables in Zeka et al. (2006a). The estimates with next to them are significantly higher
than the lowest estimate in the group.
December 2008
6-212
DRAFT-DO NOT QUOTE OR CITE
-------
Location of death
in
OJ
CE
LU
&
ui
o
\
° tn-hospital
• Out-of-hospital
&¦
o
c\i
ac
-U
Season
Winter
a Summer
Spring/Fall
~
0}
CC
0)
OC
Figure 6-17. PM10 risk estimates (per 10 pg/m3) by location of death and by season. The risk
estimates and 95% confidence intervals were plotted using numerical results from
tables in Zeka et al. (2006a). The estimates with next to them are significantly higher
than the lowest estimate in the group.
Overall, Zeka et al. (2006a) showed a consistent pattern of effect modification by contributing
causes of death (i.e., pneumonia, stroke, heart failure, and diabetes) on PMi0 risk estimates for primary
causes of death (Figure 6-18; not all results for contributing cause are shown). However, because the
contributing causes of death counts were relatively small, as reflected in the wide confidence bands in
Figure 6-18, most of the contrasts observed did not achieve statistical significance.
December 2008
6-213
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
£ 4
5
o
E
3 ¦ ¦
4) '
Ifl
nj
flj
c 14
o 1T
c
£
S
//
* ¦
# #
£ £
All cause
ii
£
11
+ '
H
J /
+ '
11
# I
.5 ^
5 Q
+ J
.£? *5
//
t *
Respiratory Ml
Primary cause of death
+ '
j® ¦$
II
if ^
+
3 ^
Stroke
Source: Zeka et al. (2006a)
Figure 6-18. PM10 risk estimates (per 10 pg/m3) by contributing causes of deaths. The estimates
with (added to the original figure) indicates a significant difference.
In addition, when examining the other effect modifiers, the results that show no gender or race
differences in PM10 risk estimates for all-cause and cardiovascular deaths are important, given the
relatively narrow confidence bands of these estimates. The effect modification by the location of death
has been reported previously in smaller studies, but the large contrast found for all-cause and
cardiovascular mortality in this large multicity analysis is noteworthy. The elevated PMi0 risks reported
by Zeka et al. (2006a) for all-cause, heart disease (and stroke) ""out-of-hospital" deaths are also consistent
with the hypothesis of acute PMi0 effects on "sudden deaths" brought on by systemic inflammation or
dysregulation of the autonomic nervous system. The finding regarding the seasonal effect modification,
though significant only for respiratory deaths, is somewhat in contrast with the Peng et al. (2005) analysis
of the extended NMMAPS data, which observed the greatest effects during the summer season. The
apparent inconsistency may be due to the difference in geographic coverage (i.e., 20 vs. 100 cities) or
methodology (i.e., case-crossover with referent days in the same month of the same year vs. time-series
analysis with adjustment for temporal trend in the regression model).
Summary of PMw Risk Estimates
Of those studies discussed in the text, depending on the lag/averaging time and the number of cities
included, the estimates for all-cause (non-accidental) mortality for all ages ranged from 0.12% (Dominici
December 2008
6-214
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
et al., 2007b) to 0.53% (Schwartz, 2004b) per 10 (ig/m3 increase in PM10, regardless of the study design
used (i.e., time-series vs. case crossover) (see Figure 6-19). The majority of studies examined present
estimates for either a lag of 1 day or a 2-day average (lag 0-1), both of which have been found to be
strongly associated with the risk of death (Schwartz, 2004b, c). However, the use of a distributed lag
model was found to result in slightly larger (by -30%) estimates compared to those for single-day lags
(Zeka et al., 2005). Overall, an examination of the PMi0 risk estimates stratified by cause-specific
mortality and age, for all U.S.- and Canadian-based studies, further supports the findings of the multicity
studies discussed in the 2004 PM AQCD and this ISA, but there is a larger degree of variability and
uncertainty in risk estimates derived from single-city studies (see Figure 6-20).
The variability in PMi0 mortality risk estimates reported within and between multicity studies may
be due to the difference in the cities analyzed and the potential regional differences in PM composition.
The NMMAPS studies have found that geographic regions and seasons are the two most important factors
that determine the variability in risk estimates, with estimates being larger in the Eastern U.S. and during
the summer, respectively. These findings were fairly consistent across studies, but Zeka et al. (2006a) did
observe the strongest association during the transition period (spring and fall); however, this may be due
to the difference in geographic coverage or the difference in the model specification used compared to
Peng et al. (2005).
Finally, examination of potential confounders showed that the size of PM10 risk estimates are fairly
robust to the inclusion of gaseous copollutants in models (Peng et al., 2005) or by matching days with
similar gaseous pollutant concentrations (Schwartz, 2004b). These findings further confirmed that PM10
risk estimates are not, at least in a straightforward manner, confounded by gaseous copollutants.
December 2008
6-215
DRAFT-DO NOT QUOTE OR CITE
-------
Peng et al. (2005), NMMAPS, 1
All seasons
i
i
i
1
[ •
Winter
Spring
Summer
Fall
1
1 •
1
j ~
1 *
j 9
Domlnlcl et al. (2007), NMMAPS, 1
1
Nationwide, 1987-1994
1995-2000
1987-2000
[ •
I •
1
East, 1987-1994
1995-2000
1987-2000
~
1 •
l
1
West, 1987-1994
1995-2000
1987-2000
j •
1 *
-1 •
Welty & Zeger (2005), NMMAPS, 1
Schwartz (2004a), 14 U.S. cities, 1
i
1
1 mini
¦
Bidirectional, 2-stage
Bidirectional. 1-stage
Matched by temperature, 2-stage
Matched by temperature, 1-stage
Time-series
•
i •
1 •
i
, •
i •
Schwartz (2004b), 14 U.S. cities, 0-1
i
i
Matched by CO (13 cities)
Matched by 03 (13 cities)
Matched by N02 (8 cities)
Matched by S02 (10 cities)
1 •
| 9
l •
' •
Zeka et al. (2005), 20 U.S. cities
I
0
1
2
Sum of distributed lag 0-2
[ •
l •
1 •
l
, »
Zeka et al. (2006), 20 U.S. cities, 1-2
I
i
Winter
Summer
Spring/Fall
i
• •
i •
1
1
In-hospital
Out-of-hospital
i
, •
' •
1
1
l I l 1 i I
-0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2
% Increase
Figure 6-19. Summary of PM10 risk estimates (per 10 pg/m3) for all-cause mortality from recent
multicity studies. The number after the study location indicates lag/average used for
PM10 (e.g., "01" indicates the average of lag 0 and 1 days). For Welty and Zeger (2005),
the vertical lines represent point estimates for 23 different weather models, and the
horizontal band spans the 95% posterior intervals of these point estimates.
December 2008
6-216
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Location
Lag
20MPMAQCD
Dominici etal. (2007a)
Peng et at (2005)
Roberts ami Martin (2007a)
Welly and Zeger (2005)
Dcminicietal. (2008a)
Zeka et at. (2005)*
Daniels etal. (2004)
Schwartz (2004a)
Burnett et al. (2004)
Dominici et al. (2004b)
Bateson and Schwartz (2004)
Slaughter et al. (2005)
Wetyetal (2008)*
ZeJra et al. (2006)
Zeka et al. (2006)
Zeka et al. (2006)
Villeneuveet al. (20Q3)
Roberts (2004)
Roberts (2004)
2004 PM AQCD
Zeka et al. (2005)*
Villenaiveet al. (2003)
DeLeon et al. (2003)
DeLeon et al. (2003)
DeLeon et al. (2003)
Zeka et al. (2005)*
Zeka et al. (2006)
Zeka el al. (2006)
Zeka et al. (2006)
Zeka et al. (2005)*
Zeka et al. (2006)
Zeka et al. (2006)
Zeka et al. (2006)
Zeka etal. (2005)*
Zeka et al. (2005)*
Zeka et al. (2005)
Zeka et al. (2006)
Zeka et al. (2006)
Zeka et al. (2006)
20G4PMAQCD
Zeka et al. (2005)*
Zeka et al. (2006)
Zeka etal. (2006)
Zeka et al. (20D6)
Villeneuveet al. (2003)
Zeka et al. (2005)*
Zeka el al. (2005)*
Dcminicietal. (2007a)
Roberts arid Martin (2007a)
Dcminicietal. (2003a)
Daniels etal. (2004)
Dominici etal. (2007a)
Dominici et si. (2003a)
Daniels et al. (2004)
U.S. & Canada
100U.S, cities
100U.S. cities
100 U.S. cities
100 U.S. cities
68 U.S. cities
20 U.S. cities
20 U.S. dies
14 U.S. cities
12 Canadian cities
10 U.S. cities
Cook County, Illinois
Spokane. Washington
Chicago. Illinois
20 U.S. cities
20 U.S. dies
20 U.S. dies
Vancouver, Canada
Cook County, Illinois
1
1
0-2
1
1
3-day cum
0-1
1
1
0-1
0-1
1
0-14
1-2
1-2
1-2
0-2
0-3
Allegheny County, Pennsylvania 0-3
U.S. & Canada
20 U.S. dies
Vancouver, Canada
New York. MY
Mew York, NY
New York, MY
20 U.S. cities
20 U.S. cities
20 U.S. dies
20 U.S. cities
20 U.S. dies
20 U.S. cities
20 U.S. cities
20 U.S. cities
20 U.S. cities
20 U.S. cities
20 U.S. cities
20 U.S. dies
20 U.S. cities
20 U.S. dies
U.S. & Canada
20 U.S. cities
20 U.S. cities
20 U.S. cities
20 U.S. cities
Vancouver, Canada
20 U.S. dies
20 U.S. dies
100 U.S. cities
100 U.S. cities
88 U.S. cities
20 U.S. dies
100 U.S. cities
88 U.S. cities
20 U.S. cities
3-day cum
0
0-1
0-1
0-1
3-day cum
2
2
2
3-day cum
0
0
0
3-day cum
3-day cum
1
1
1
1
3-day cum
0-3
0-3
0-3
0
3-day cum
3-day cum
1
0-2
1
0-1
!
1
0-1
Non-acddental
Cardiovascular
Respiratory
Cardio-Respiratory
Other-causes
-65-75
.>75
->75
75J Circulatory
- IHD
-65-75
->75 J
Heart Disease
->75
— 65-75
— HF
. Dysrhythmias
]'
65-75
. >75 J
Stroke
.65-75
.>75
—COPD
Pneumonia
-2
T
-1
r
T
0 1
% Increase
Figure 6-20. Summary of PM10 risk estimates (per 10 |jg/m3) for cause-specific mortality for all U.S.
and Canadian-based studies. The estimates provided for "2004 PM AQCD" represent the
lowest and highest central estimates for the U.S.- and Canadian-based studies evaluated in the previous
AQCD. The "x" presesented in the non-accidental mortality range represents the lone multicity study
evaluated (Dominici et al., 2003a). For Welty and Zeger (2005) the vertical lines represent point estimates for
23 different weather models, and the horizontal band spans the 95% posterior intervals of these point
estimates. Circle: all ages; triangle: < 65; square > 65; *:- distributed lag model.
December 2008
6-217
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
6.5.2.2. PM2.5
Nationwide collection of PM2 5 data began in 1999. This in conjunction with the unavailability of
nationwide mortality data from the NCHS starting with year 2001 data, as discussed previously, resulted
in only a few multicity studies (i.e., studies in which mortality data was obtained from state or city
agencies), which could examine the mortality effects of PM25 for an extended period of time.
PM2.5 - Mortality Associations in 100 U.S. cities, 1999 to 2000
The Dominici et al. (2007b) NMMAPS study (described in Section 6.5.2.1.), also examined
PM2 5-mortality associations using the same methodology and data for 1999-2000. The PM2 5 risk
estimates at lag 1-day were 0.29% (PI: 0.01, 0.57) and 0.38% (PI: -0.07, 0.82) per 10 (ig/m3 increase for
all-cause and cardio-respiratory mortality, respectively. The authors also conducted a sensitivity analysis
of the risk estimates based on the extent of adjustment for temporal trends in the model, changing the
degrees of freedom (df) of temporal adjustment from 1 to 20/yr (the main result used 7 df/yr). In
comparison to the PMi0 results, the PM2 5 risk estimates appeared more sensitive to the extent of temporal
adjustment between 5 and 10 df/yr, but this may be in part due to the much smaller sample size used for
the PM2 5 analysis compared to the PMi0 analysis.
PM2.5 - Mortality Associations in 27 U.S. Cities, Variable between 1997 and 2002
Franklin et al. (2007) analyzed 27 cities that had PM2 5 monitoring and daily mortality data for at
least 2 years of a 6-year period 1997 to 2002. The mortality data up to year 2000 were obtained from the
NCHS, while the 2001-2002 data were obtained from six states (CA, MI, MN, PA, TX, and WA),
resulting in 12 out of the 27 cities having data up to 2002. The start year for each city included in the
study was set at 1999, except for Milwaukee, WI (1997) and Boston, MA (1998), which is due to PM2 5
data availability in these two cities. In the case-crossover analysis in each city, control days for each death
were chosen to be every third day within the same month and year that death occurred in order to reduce
auto-correlation. The first stage regression examined the interaction of effects with age and gender, while
the second stage random effects model combined city-specific PM2 5 risk estimates and examined possible
effect modifiers using city-specific characteristics (e.g., prevalence of central air conditioning and
geographic region). For all of the mortality categories, the estimates for lag 1-day showed the largest
estimates. The combined estimates at lag 1 day were: 1.2% (CI: 0.29, 2.1), 0.94% (CI: -0.14, 2.0), 1.8%
(CI: 0.20,3.4), and 1. 0% (CIi 0.02, 2.0) for cill-Cciuse, Ccirdiovcisciilcir, respiratory, ctnd stroke deaths,
respectively, per 10 (ig/m3. When examining the city-specific risk estimates most of the cities with
negative estimates are also those with a high prevalence of central air conditioning (Dallas, 89%;
December 2008
6-218
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Houston, 84%; Las Vegas, 93%; Birmingham, 77%). It is unclear why these cities exhibit negative (and
significant) risk estimates rather than null effects.
In the analysis of effect modifiers, Franklin et al. (2007) found that individuals > 75 showed
significantly higher PM2.5 risk estimates. The estimated effects were also found to vary by geographic
location with larger estimates in the East than in the West, which are consistent with the regional pattern
found in the NMMAPS PMi0 risk estimates. In addition, a higher prevalence of central air conditioning
was associated with decreased PM2 5 risk estimates when comparing the lower (25th percentile) vs. the
higher (75th percentile) air conditioning use rates, especially in the cities where PM2 5 concentrations peak
in the summer. Finally, the risk estimates were not found to be different between communities with PM2 5
levels less than or equal to vs. higher than 15 (ig/m3. The risk estimates for each effect modifier are
presented in Figure 6-21. Note the wide confidence intervals associated with each of the risk estimates,
specifically for Franklin et al. (2007) and Ostro et al. (2006), which suggests low statistical power for
testing the differences between effect modifiers.
PM2.5 - Mortality Associations in 25 U.S. Cities between 2000 and 2005
Franklin et al. (2007) analyzed 25 cities that had PM2 5 monitoring and daily mortality data between
the years 2000 to 2005 (with the study period varying from city to city). The choice of the 25
communities was based on the availability of PM2 5 mass concentrations and daily mortality records for at
least 4 years, along with PM2 5 speciation data for at least two years between 2000 and 2005. Similar to
Franklin et al. (2007), all-cause, cardiovascular, respiratory, and stroke deaths were examined; however,
of the 25 cities included in the study, only 15 overlap with the 27 cities analyzed in Franklin et al. (2007).
The authors obtained mortality data from the NCHS and various state health departments (California,
Massachusetts, Michigan, Minnesota, Missouri, Ohio, Pennsylvania, Texas, and Washinton). Although the
main objective of the study was to examine the role of PM25 chemical species in the second stage
analysis, the first stage analysis conducted a time-series regression of mortality on PM2 5. In addition, the
first stage regression performed a seasonal analysis in order to take advantage of seasonal variation in
PM2 5 chemical species across cities and to possibly explain the city-to-city variation in PM2 5 mortality
risk estimates. From this analysis a strong seasonal pattern was observed with the greatest effects
occurring in the spring and summer seasons (see Figure 6-21).
Overall, the risk estimates for all-cause, cardiovascular, and respiratory deaths reported by Franklin
et al. (2008) are comparable to those presented in the 27 cities study (Franklin et al., 2007) and the
California 9 counties study (Ostro et al., 2006), as shown in Figure 6-21. When comparing the 2007 and
2008 studies conducted by Franklin et al., although only 15 cities overlap between the two studies and
each study was designed differently (i.e., time-series vs. case-crossover), the magnitude of the PM2 5 risk
December 2008
6-219
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
estimates reported were similar for the same averaging time, and both studies reported a regional pattern
(East > West) similar to that found in NMMAPS for PM10 risk estimates.
PM2.5 - Mortality Associations in Nine California counties, 1999-2002
Ostro et al. (2006) examined associations between PM2 5 and daily mortality in nine heavily
populated California counties (Contra Costa, Fresno, Kern, Los Angeles, Orange, Riverside, Sacramento,
San Diego, and Santa Clara) using data from 1999 through 2002. The authors used a two-stage model to
examine all-cause, respiratory, cardiovascular, ischemic heart disease, and diabetes mortality individually
and by potential effect modifier (i.e., age, gender, race, ethnicity, and education level). The a priori
exposure periods examined included the average of 0- and 1-day lags (lag 0-1) and the 2-day lag (lag 2).
The authors selected these non-overlapping lags (i.e., rather than selecting lag 1 as the single-day lag)
because previous studies have reported stronger associations at lags of 1 or 2 days or with cumulative
exposure over three days. It is unclear why the investigators chose these non-overlapping lags
(i.e., single-day lag of 2 instead of 1) even though they state they based the selection of their lag days on
results presented in previous studies, which found the strongest association for PM lagged 1 or 2 days.
Using the average of 0- and 1-day lags Ostro et al. (2006) reported combined estimates of: 0.6% (CI: 0.2,
1.0), 0.6% (CI: 0.0, 1.1), 0.3% (CI: -0.5, 1.0), 2.2% (CI: 0.6, 3.9), and 2.4% (CI: 0.6, 4.2) for all-cause,
cardiovascular, ischemic heart disease, respiratory, and diabetes deaths, respectively, per 10 (ig/m3. The
risk estimates for the major underlying causes and for potential effect modifiers are presented in Figures
6-21 and 6-22. The authors also conducted a sensitivity analysis of risk estimates based on the extent of
temporal adjustment, which showed monotonic reductions for all of the death categories examined when
4, 8, and 12 degrees of freedom per year were used.
Five of the nine counties examined in the Ostro et al. (2006) analysis are among the 27 cities
examined in the Franklin et al. (2007) analysis for the same period, 1999-2002. While the lags used were
different between these two studies, both presented PM2.5 risk estimates in individual cities (graphically in
the Franklin et al. study; in a table in the Ostro et al. study), which allowed for a cursory evaluation of
consistency between the two analyses. In Franklin et al. (2007) PM2 5 risk estimates at lag 1 day for Los
Angeles and Riverside were slightly negative, whereas Fresno, Sacramento, and San Diego showed
positive values above 1% per 10 (ig/m3 increase in PM2 5. The 2-day lag result presented in Ostro et al.
(2006) is qualitatively consistent, with Los Angeles and Riverside, both of which show slightly negative
estimates, while the other 3 locations all show positive, but somewhat smaller estimates, than those
reported by Franklin et al (2007). The estimates for the average of 0- and 1-day lags for these five cities in
Ostro et al. (2006) were all positive. Thus, these two PM2 5 studies showed some consistencies in risk
estimates even though they used different lag periods. Although the risk estimates for Franklin et al.
December 2008
6-220
DRAFT-DO NOT QUOTE OR CITE
-------
1 (2007) and Ostro et al. (2006), stratified by various effect modifiers (gender, race, etc.), are summarized
2 in Figure 6-21 there is one noteworthy contrast. Ostro et al. (2006) observed comparable risk estimates for
3 ""in-hospital" vs. ""out-of-hospital" deaths, which is in contrast to the large difference between the two
4 found for risk estimates in the 20 cities study discussed earlier (Zeka et al., 2006a).
Ostro et al. (2006), 9 CA Counties, 0-1
All-cause
Age > 65
•
•
Male
Female
•
White
Black
Hispanic
*
•
In-hospital
Out-of-hospital
•
> High School
< High School
•
Franklin et al. (2007), 27 U.S. cities, 1
All-cause
•
Age a 75
Age < 75
•
Male
Female
•
East
West
•
PM2.5 > 15pg/m3
PM2.5 5 15|jg/m3
25th percentile air conditioning
75th percentile air conditioning
Summer peaking PM2.5 cities:
25th percentile air conditioning
75th percentile air conditioning
Franklin et al. (2008), 25 U.S. cities, 0-1
All-cause
•
Winter
Spnng
Summer
Fall
r-m
•
!—•
West
East & Central
m
•
I I I
-10 1 2 3
% Increase
Figure 6-21. Summary of all-cause mortality PM2.5 risk estimates per 10 pg/m3 by various effect
modifiers.
December 2008
6-221
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Multi-city Studies on PM2.5 Mortality Effects in Other Countries
Burnett et al.'s study of multiple pollutants in 12 Canadian cities found the most consistent
associations for N02 (2004). In this analysis, PM2 5 was only measured every sixth day in much of the
study period, and a simultaneous inclusion of N02 and PM2 5 in a model on the days when PM2 5 data
were available eliminated the PM2 5 association (from 0.60% to -0.10% per 10 (ig/m3 increase in PM2 5).
However, the investigators noted that during the later study period of 1998-2000 when daily TEOM PM2 5
data were available for 11 of the 12 cities, a simultaneous inclusion of N02 and PM2 5 resulted in
considerable reduction of the N02 risk estimate, while the PM2 5 risk estimate was only slightly reduced
from 1.1% to 0.98% (CI: -0.16, 2.14). Thus, the relative importance ofN02 and PM2 5 on mortality effect
estimates has not been resolved when using the Canadian data sets.
Summary of PM2.5 Risk Estimates
The risk estimates for all-cause mortality for all ages ranged from 0.29% (Dominici et al., 2007) to
1.21% (Franklin et al., 2007) per 10 (ig/m3 increase in PM2 5 (see Figure 6-22). An examination of
cause-specific risk estimates found that PMi0 risk estimates for cardiovascular deaths are similar to those
for all-cause deaths (0.30-1.03%), while the effect estimates for respiratory deaths were consistently
larger (1.01-2.2%), albeit with larger confidence intervals as well, than those for all-cause or
cardiovascular deaths using the same lag/averaging indices. Figure 6-23 summarizes the PM2 5 risk
estimates for all U.S.- and Canadian-based studies by cause-specific mortality and age.
An examination of lag structure observed results similar to those reported for PMi0 with most
studies reporting either single day lags or two-day average lags with the strongest effects observed on lag
1 or lag 0-1. In addition, seasonal and regional patterns of PM2 5 risk estimates were found to consistently
support those reported for PM10, with the warmer season and Eastern U.S. showing the strongest
association. However, unlike the examination of PM10 risk estimates, no U.S.-based multicity studies
analyzed potential confounding of PM2 5 risk estimates by gaseous pollutants. Burnett et al. (2004) in a
Canadian multicity study did analyze gaseous pollutants and found mixed results, with possible
confounding by N02. Therefore, it is unclear if gaseous pollutants confound the PM2 5 mortality
association.
December 2008
6-222
DRAFT-DO NOT QUOTE OR CITE
-------
Dominici et al. (2007), NMMAPS
All-cause, lag 1
Cardiorespiratory, lag 1
—•—
—•
Franklin et al. (2007), 27 U.S. cities
All-cause, lag 1
Cardiovascular, lag 1
Respiratory, lag 1
•
•
~
All-cause, lag 0-1
Cardiovascular, lag 0-1
Respiratory, lag 0-1
•
•
•
Franklin et al. (2008), 25 U.S. cities
All-cause, lag 0-1
Cardiovascular, lag 0-1
Respiratory, lag 1-2
—•
~
~
Ostro et al. (2006), 9 CA counties
All-cause, lag 0-1
Cardiovascular, lag 0-1
Respiratory, lag 0-1
•
*
1 I I 1
-10 12 3 4
% Increase
Figure 6-22. Summary of PM2.5 risk estimates per 10 pg/m3 for major underlying causes of death.
December 2008
6-223
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Location
Lag
2004 PM AQCD
U.S. & Canada
-x—•
Non-accidental
Dominicietal. (2007a)
100 J.S. cities
1
r«-
Franklin el al. (2007)
27 U.S. cities
1
•
Franklin el al. (2008)
25 U.S. cities
0-1
Burnett et al. (2004)
12 Canadian cities
1
Ostroetal. (2006)
9 California counties
0-1
Slaughter et al- (2005)
Spokane, Washington
1
Klemm et al. (2004)
Atlanta, Georgia
0-1
Villeneuve et al. (2003)
Vancouver, Canada
0
2004 PM AQCD
U.S. & Canada
caraiovascuiar
Franklin et al. (2007)
27 U.S. cities
1
Franklin et al. (2008)
25 U.S. cities
0-1
Ostroetal. (2007)
9 California counties
3
Ostroetal. (2006)
Holloman et al. (2004)
Wilson et al. (2007)
Wilson et al. (2007)
Wilson et al. (2007)
9 California counties
7 NC counties
Phoenix, Arizona
Phoenix, Arizona
Phoenix, Arizona
0-1
0
0-5
0-5
0-5
>16
> 25; Central
>25; Middle
> 25; Outer
Villeneuve etal. (2003)
Vancouver, Canada
1
Goldberg etal. (2003)
Montreal, Quebec, Canada
0
Franklin etal. (2007)
27 U.S. cities
1
1 Stroke
2004 PM AQCD
Franklin etal. (2007)
U.S. & Canada
27 U.S. cities
Respiratory
0-1
Franklin etal, (2008)
25 U.S. cities
1-2
Ostroetal. (2006)
9 California counties
0-1
•
Villeneuve etal. (2003)
Vancouver, Canada
0
Dominicietal. (2007a)
100 U.S. cities
1
Cardiorespiratory
Goldberg etal. (2006)
Montreal, Quebec, Canada
0-2
Diabetes.
• All ages ¦>
65 yrs old
i I i i I I
-2 0 2 4 6 8 10 12
% Increase
Figure 6-23. Summary of PM2.5 risk estimates (per 10 pg/m3) for cause-specific mortality for all
U.S.- and Canadian-based studies. The estimates provided for "2004 PM AQCD"
represent the lowest and highest central estimates for the U.S.- and Canadian-based
studies evaluated in the previous AQCD. The "x" presented in the non-accidental
mortality range represents the lone multi-city study evaluated (Burnett and Goldberg,
2003).
December 2008
6-224
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
6.5.2.3. Other Size-fractionated PM Indices
Currently, the U.S. does not have a monitoring network in place to measure size-fractionated PM
indices other than PM10 and PM2 5; as a result, no U.S.-based multicity studies have recently been
conducted that examine other PM size fractions. In the 2004 PM AQCD, there were several U.S.-based
studies that examined both fine (PM2 5) and thoracic coarse (PM^.2.5) PM for their associations with
mortality. However, since then, very few U.S.- and Canadian-based studies have examined PM10-2.5. Due
to the limited body of literature that has examined the association between short-term exposure to PMi 0-2.5
and mortality, unlike previous sections which focused specifically on U.S.- and Canadian-based studies,
this section will review single-city studies and those studies conducted in other countries that have PMi0_
2 5 concentrations similar to those found in the U.S. and Canada. Due to the varying model specifications
and lags examined in these studies, quantitative synthesis of the risk estimates requires caution.
Thoracic Coarse Particles (PM10-2.5)
While a large percent of the composition of coarse particles may consist of crustal materials by
mass, depending on available sources, the surface chemical characteristics of PMi0-2.5 may also vary from
city to city. Thus, without information on the chemical speciation of PM10-2.5, the apparent variability in
observed associations between PM10-2.5 and mortality across cities is difficult to characterize.
Nevertheless, the relative importance of the associations observed between PMi0_2.5 and mortality in the
following studies is of interest.
In Burnett et al. (2004), which analyzed the association of multiple pollutants with mortality in
12 Canadian cities, described previously, the authors also examined PMi0.2.5. In this study the authors
collected PMi0_2.5 using dichotomous samplers with an every-6th-day schedule. When both N02 and
PM10-2.5 were included in the regression model, the PM10.2.5 effect estimate was reduced from 0.65% (CI:
-0.10, 1.4) to 0.31% (95% CI: -0.49 to 1.1) per 10 (ig/m3 increase in 1-day lag PM10.2 5. These risk
estimates are similar to those reported for PM2 5, which were also reduced upon the inclusion of N02 in
the two-pollutant model, but to a greater extent from 0.60% (95% [CI: -0.03 to 1.2]) to -0.1% (95% [CI:
-0.86 to 0.67]).
Villeneuve et al. (2003) analyzed the association between PM2 5, PM10-2.5, TSP, PMi0, sulfate, and
gaseous copollutants in Vancouver, Canada, using a cohort of approximately 550,000 whose vital status
was ascertained between 1986 and 1999. The authors examined each air pollutant's association with
all-cause, cardiovascular, and respiratory mortality, but only observed significant results for
cardiovascular mortality at lag 0 for both PMi0_2.5 and PM2 5. They found that PMi0_2.5, (5.4% [95% CI:
1.1, 9.8] per 10 |_ig/m3 '. was more strongly associated with cardiovascular mortality than PM2 5, (4.8%
[95% CI: -1.9 to 12.0] per 10 (ig/m3).
December 2008
6-225
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Klemm et al. (2004) analyzed various components of PM and gaseous pollutants for their
associations with mortality in Fulton and DeKalb Counties, Georgia for the two-year period, 1998-2000.
In this analysis the authors adjusted for temporal trend using quarterly, monthly, and biweekly knots, and
reported estimates for all-cause, circulatory, respiratory, cancer, and other causes mortality for each
scenario. Overall, PM2.5 was, generally, more strongly associated with mortality than PMi0_2.5. For
example, using the average of 0- and 1-day lags, the risk estimates for PM2 5 and PM10-2.5 in the monthly
knots model for all-cause mortality, ages > 65 were 5.6% (95% [CP 1.9, 9.5]) and 6.4% (95% [CP -0.5,
14.1]) per 10 (ig/m3 increase, respectively.1
Slaughter et al. (2005) examined the association of various PM size fractions (PMi, PM2 5, PMi0,
PM10-2.5) and CO with ED visits, HAs, and mortality in Spokane, WA for the period 1995-2001. Although
the authors did not report mortality risk estimates for PM10-2.5, they did not find an association with any
PM size fraction (or CO) with mortality or cardiac HAs at the 0- to 3-day lag.
Wilson et al. (2007b) examined the association between size-fractionated PM (PM2 5 and PM10-2.5)
and cardiovascular mortality in Phoenix for the study period 1995-1997, using mortality data aggregated
for three geographic regions: "Central Phoenix," "Middle Ring," and "Outer Phoenix," which were
constructed as a composite of ZIP codes of residence in order to compare population size among the three
areas. The authors reported apparently different patterns of associations between PM2 5 and PM10_2.5 in
terms of the size of the risk estimate across the three areas and temporal patterns of associations. In the
"Middle Ring" where PM10_2.5 showed the strongest association, the estimated risk per 10 (ig/m3 increase
for a 1 day lag was 3.4% (95% CI: 1.0-5.8). The estimated risk for PM2 5 found for "Central Phoenix" was
6.6% (95% CI: 1.1-12.5) for lag 1. The authors speculated that the apparent difference in estimated risks
across the areas might be due to the lower SES in "Central Phoenix" or the lower exposure error, but the
relatively wide confidence bands of these estimates make it difficult to establish such relationships.
Kettunen et al. (2007) analyzed ultra-fine particles, PM2 5, PMi0, PM10-2.5, and gaseous pollutants
for their associations with stroke mortality in Helsinki during the study period of 1998-2004. The authors
did not observe an association between air pollution and mortality for the whole year or cold season, but
they did find associations for PM2.5 (13.3% [95% CI: 2.3-25.5] per 10 (.ig/ni3). PMi0, and CO during the
warm season, most strongly at lag 1 day. An association was also observed for PMi0_2.5 during the warm
season (7.8% [95% CI: -7.4-25.5] per 10 (ig/m3 at lag 1 day); however, it was weaker than PM2 5.
1 The monthly knot model was selected for comparison because, overall, PM2 5 showed the strongest association with all-cause mortality among
the 15 air pollution indices examined when using this model.
December 2008
6-226
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
The Perez et al. (2008) analysis tested the hypothesis that outbreaks of Saharan dust exacerbate the
effects of PM2 5 and PM10.2.5 on daily mortality. Changes of effects between Saharan and non-Saharan dust
days were assessed using a time-stratified case-crossover design involving 24,850 deaths between March
2003 and December 2004 in Barcelona, Spain. Saharan dust days were identified from back-trajectory
and satellite images. Chemical speciation, but not an analysis for microbes or fungi, was conducted
approximately once a week during the study period. On Saharan dust days, mean concentrations were 1.2
times higher for PM2.5 (29.9 |_ig/m3) and 1.1 times higher for PMi0.2.5 (16.4 (.ig/nr1) than on non-Saharan
dust days. During Saharan dust days (90 days out of 602), the PMi0.2.5 risk estimate was 8.4% (95% [CI:
1.5-15.8]) per 10 (ig/m3 increase at lag 1 day, compared with 1.4% (95% CI: -0.8% to 3.4%]) during
non-Saharan dust days. In contrast, there was not an additional increased risk of daily mortality for PM2 5
during Saharan dust days (5.0% [95% CI: 0.5-9.7]) compared with non-Saharan dust days (3.5%
[95% CI: 1.6-5.5]). However, differences in chemical composition (i.e., PM25 primarily composed of
nonmineral carbon and secondary aerosols; whereas, PMi0.2.5 dominated by crustal elements) did not
explain these observations. Note also when examining all days combined, both size fractions were
associated with mortality, but the PM2 5 association was found to be stronger.
Summary of PM10-2.5 Risk Estimates
The results from newly available studies that examined the association between short-term
exposure to PMi0.2.5 and mortality are mixed, as was the case in the 2004 PM AQCD. Due to the
relatively wide confidence bands for the mortality risk estimates from the single-city studies evaluated
along with the city-to-city variation in the chemical components of PM10.25, a quantitative summary of
PMio_2.5 effects may not be informative at this point. In addition, the mortality risk estimates associated
with PM10.2.5 may also be influenced by effect modifying conditions (e.g., season, relative exposure error,
and dust storms), and to date have not been extensively examined. Clearly, more data are needed to
characterize the chemical and biological components that may modify the potential toxicity of coarse
particles. Figure 6-24 summarizes the PMi0.2.5 risk estimates for all U.S.-, Canadian-, and international-
based studies by cause-specific mortality and age.
December 2008
6-227
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Location
• All ages
¦ > 65 yis old
Lag
2004 PM AQCD
U.S.S Canada
Burnett et al. (2004)
12 Canadian cities
1
Perez et al. (2008)
Barcelona, Spain
1
Perez et al. (2008)
Barcelona, Spain
1
Klemm et al. (2004)
Atlanta, Georgia
0-1
Villeneuve et al. (2003)
Vancouver, Canada
0
2004 PM AQCD
U.S. & Canada
Wilson et al. (2007)
Phoenix, Arizona
0-5 ma
Wilson et al. (2007)
Phoenix, Arizona
0-5 ma
Wilson etal. (2007)
Phoenix, Arizona
0-5 ma
Villeneuve et al. (2003)
Vancouver, Canada
1
Kettunen et al. (2007)
Helsinki, Finland
1
2004 PM AQCD
U.S. & Canada
Villeneuve et al. (2003)
Vancouver, Canada
1
Saharan dusf days
-2
Non-accidental
. Non-Sail a ran dust days
Cardiovascular
>25; Central
>25: Middle
. >25 Outer
Warm season
Respiratory
»
t 1 r
4 6 8
% Increase
T-
10
12
Figure 6-24. Summary of PM10-2.5 risk estimates (per 10 pg/m3) for cause-specific mortality for all
U.S.-, Canadian-, and international-based studies. The estimates provided for "2004
PM AQCD" represent the lowest and highest central estimates for the U.S.- and
Canadian-based studies evaluated in the previous AQCD. The "x" presented in the
non-accidental mortality range represents the lone multi-city study evaluated (Burnett
and Goldberg, 2003).
6.5.2.4. Ultrafine Particles
1 The 2004 PM AQCD reviewed Wichmann et al.'s (re analyzed by Stolzel et al., 2003; 2000) study
2 of fine and ultrafine particles (UFP) (diameter: 0.01-0.1 (.tin) in Erfurt, Germany for the study period
December 2008
6-228
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
1995-1998. Stolzel et al. (2007) extended the study period to include the years 1995-2001 and updated the
analysis. Number concentrations (NC) for four size ranges of UFP (0.01-0.1, 0.01-0.03, 0.03-0.05, and
0.05-0.1 |_im) as well as mass concentration (MC) for three size ranges (0.01-2.5, 0.1-0.5, and 10 |_im)
were analyzed. They found associations with UFP NC and all-cause as well as cardiorespiratory
mortality, each for a 4 day lag. The risk estimates associated with a 9,748/cm3 increase in UFP NC was
2.9% (95% CI: 0.3-5.5) for all-cause mortality and 3.1% (95% CI: 0.3-6.0) for cardiorespiratory
mortality. The UFP-mortality association, and the lag structure of association, is consistent with the
results from their earlier analysis, but the PM2 5 association found in the previous study was not observed
in the updated analysis. Both UFP and PM2 5 concentrations were higher during the cold season in this
locale.
Kettunen et al.'s (2007) study in Helsinki also examined the relationship between UFP and stroke
mortality. As described earlier, PM2 5, PMi0, and CO was associated with stroke mortality only during the
warm season. The association with UFP was borderline non-significant (8.5% [95% CI: -1.2 to 19.1] per
4,979/cm3 increase in UFP at lag 1 day), but its lag structure of association and the magnitude of the
effect estimate per interquartile-range are similar to those for PM2 5. Note that the UFP NC levels in
Helsinki (median equals 8,986/cm3 during the cold season and 7,587/cm3 during the warm season) are
lower than those in Erfurt (mean = 13,549/cm3), but clearly higher in the cold season.
Summary of Ultra-Fine Particle Risk Estimates
Only two new studies reported associations between UFP and mortality. In Erfurt, UFP showed the
strongest associations with mortality among all of the PM indices, but its lag structure of association is
unique (strongest association at lag 4 days) and not consistent with the lag structures of mortality found
for other PM indices in past studies. In Helsinki, the association between UFP and stroke mortality was
weaker than that for PM2 5, but its lag structure of association was similar to that for PM2 5 (strongest at
lag 1 day). However, Kettunen et al. (2007) only examined lags 0 through 3 days. Clearly, more research
is needed to further investigate the role of UFP on PM-mortality associations.
6.5.2.5. Chemical Components of PM
To date, there have only been a few studies that examined the association between mortality and
components of PM2 5. This endeavor has been undertaken by some investigators through the use of the
newly available PM2 5 chemical speciation network data. The PM2 5 chemical speciation network consists
of more than 250 monitors that have been collecting over 40 chemical species since 2000; however, most
sites started collecting data in 2001. One caveat to the new network is that because the sampling
frequencies of the monitors are either every third day or every sixth day, there have not been, generally, a
December 2008
6-229
DRAFT-DO NOT QUOTE OR CITE
-------
sufficient number of days to examine associations with mortality in single cities. To circumvent this issue,
some investigators (Dominici et al., 2007a; Franklin et al., 2008; Lippmann et al., 2006) have used the
PM2 5 chemical species data in a second stage regression to explain heterogeneity of PM10 or PM2 5
mortality risk estimates across cities. However, there have been some studies that directly analyzed PM2 5
data (e.g., Klemm et al., 2004 and Ostro et al., 2007).
Ni-
V-
EC-
Zn-
S0,-
Cu-
Pb-
oc-
pm15-
Se-
Cr-
Mn-
Fe-
As-
N03-
Al-
Si-
Source: Lippmann et al. (2006)
Figure 6-25. Percent increase in PM10 risk estimates (point estimates and 95% confidence
intervals) associated with a 5th-to-95th percentile: increase in PM2.5 and PM2.5
chemical components. The PM2.5 chemical components were log-transformed in the
regression. The PM10 risk estimates were for 60 NMMAP cities for 1987-1994.
Explaining Heterogeneity of PM10 Risk Estimates Using PM2.5 Chemical
Speciation Data in the Second Stage Regression
Lippmann et al. (2006), in addition to their primary analysis1, investigated the consistency of the
associations between specific elements and health outcomes by examining the heterogeneity of published
1 The main focus of the study was to examine the role of PM25 chemical components in a mouse model of atherosclerosis (ApoE^ ) exposed to
concentrated fine PM (CAPs) in Tuxedo, NY.
-0.5 0.0 0.5 1.0
Percent per 10-f.ig/m3 increase in PM](I
December 2008
6-230
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
1-day lagged NMMAPS PM10 mortality risk estimates for 1987-1994 across cities as a function of the
average PM2 5 chemical components across cities. They matched PM2 5 chemical species in 60 out of 90
cities. Lippmann et al. (2006) noted that the concentrations of the 16 chemical species examined averaged
over the years 2000-2003 were highly skewed across cities. They, therefore, regressed PMi0 risk estimates
on each of the PM2 5 components, raw and log-transformed, with weights based on the standard error of
the PM10 risk estimates. The log-transformed values yielded better predictive power, and, the authors
subsequently, presented the results with log-transformed values. As shown in Figure 6-25, the 16 PM2 5
species showed varying extent of predictive power in explaining the PMi0 risk estimates across 60 cities,
with nickel (Ni) and vanadium (V) being the best predictors.
No communities removed
¦o
s
o
E
05
.2
o
-s
©
New York removed
1 1—
-20 -10 0 10 20 30
Percent increase in PM10 risk estimates per IQR Ni
Source: Dominici et al. (2007a)
Figure 6-26. Sensitivity of the percent increase in PM10 risk estimates (point estimates and 95%
confidence intervals) associated with an interquartile increase in Ni. The Ni
concentration was not log-transformed in this regression model. The PM10 risk
estimates were for 72 NMMAP cities for 1987-2000. The top estimate is achieved by
including data for all the 69 communities. The other estimates are calculated by
excluding one of the 69 communities at a time.
Dominici et al. (2007a) examined the influence of Ni and V on the updated NMMAPS PMKI
mortality risk estimates for 1987-2000, using 72 counties in which Ni and V data were collected. A
Bayesian hierarchical model was used to estimate the role of Ni and V on the heterogeneity of PMm risk
estimates. While they found both Ni and V to be significant predictors of variation in PMi0 mortality risk
estimates across cities, they also noted that this result was sensitive to inclusion of the New York City
December 2008
6-231
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
data. Lippmann et al. (2006) and Dominici et al. (2007a) both reported that the Ni levels in New York
City are particularly high (-10 times the national average). Figure 6-26 shows the result of the sensitivity
analysis for Ni. Note that the Ni in this result was not log-transformed, as clearly reflected in the change
in the width of confidence bands when the New York data were removed (i.e., a skewed distribution
produces narrow bands). Dominici et al. (2007a) further noted that they reached "the same conclusion"
when log-transformed data were used in the analysis, but the results were not presented.
Explaining Heterogeneity of PM2.5 Risk Estimates Using PM2.5 Chemical
Speciation Data in a Second Stage Regression
The first stage of the Franklin et al. (2008) 25 cities study, described previously, focused on a
time-series regression of mortality on PM2 5 by season. In the second stage random effects meta
regression, the PM2 5 mortality risk estimates (25 cities x 4 seasons = 100 estimates) were regressed on the
ratio of mean seasonal PM2 5 species to the total PM2 5 mass. The authors included those species that had
at least 25% of the reported concentrations above the minimum detection limit, which resulted in 18
species being included in the analysis. Their rationale for using species proportions as effect modifiers,
according to the investigators, was that "in the first stage of the analysis the mortality risk was estimated
per unit of the total PM2 5 mass, which encompassed all measured species, and therefore it would not be
meaningful to use the species concentrations directly as the effect modifier" (Franklin et al., 2008). In the
second stage regression model, Franklin et al. also included a quadratic function of seasonally averaged
temperature to capture the inverted U-shape relationship between PM2 5 penetration and temperature.
They found that the fitted relationship between PM2 5 risk estimates across cities and seasonally averaged
temperature substantiates the use of temperature as a surrogate for ventilation (Franklin et al., 2008).
Table 6-15 shows the resulting effect modification by PM2 5 species. Al, As, Ni, Si, and sulfate were found
to be significant effect modifiers of PM2 5 risk estimates, and simultaneously including Al, Ni, and sulfate
together, or Al, Ni, and As together further increased explanatory power. Of all the species examined, Al
and Ni explained the most residual heterogeneity. Franklin et al. also examined the effect of demographic
variables on PM2 5 risk estimates and found that only median household income was significantly
associated with mortality.
December 2008
6-232
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
Table 6-15. Effect modification of composition on the estimated percent increase in mortality with
a 10 |jg/m3 increase in PM2.5-
Cause
Species
p-value for effect
modification by species to
PM2.5 mass proportion
% increase in non-accidental mortality per 10 |jg/m3
increase in PM2.5 for an interquartile increase in
species to PM2.5 mass proportion*
Heterogeneity
explained (%)f
Non-accidental
Al
<0.001
0.58
45
Univariate
As
0.02
0.55
35
Br
0.11
0.38
5
Cr
EC
Fe
K
0.12
0.79
0.33
0.06
16
0
0.43
0.12
3
Mn
0.10
0.41
28
Na+
0.42
0.14
10
Ni
0.22
0.20
14
NOs
0.01
0.37
41
NH4
0.07
-0.49
28
OC
0.84
0.04
3
Pb
Si
S042"
V
0.59
0.31
0.03
-0.02
0.17
0.41
4
11
25
Zn
0.01
0.51
33
0.28
0.30
3
0.19
0.23
15
Non-accidental
Al
<0.001
0.79
Multivariate (1)
Ni
0.01
0.34
100
S042"
<0.001
0.75
Non-accidental
Al
<0.001
0.61
Multivariate (2)
Ni
0.01
0.35
100
As
<0.001
0.58
Adjusted for temperature
tlncludes heterogeneity explained by temperature
Source: Franklin et al. (2008)
Although Lippmann et al. (2006) used NMMAPS PMi0 risk estimates and Franklin et al. (2008)
used PM2 5 risk estimates to examine effect modification due to various PM species, 14 out of the 18
species analyzed in these two studies overlap (see Figure 6-25 and Table 6-15). Both studies found that Ni
explained the heterogeneity in PM risk estimates. Note that New York City was not included in the 25
cities examined in Franklin et al. and, thus, could not influence the result. Sulfate positively, but not
significantly, explained the PMi0 risk estimates in the Lippmann et al. (2006) analysis. However, sulfate
was a significant predictor of PM2.5 risk estimates in the Franklin et al. (2008)analysis. Al and Si were
negative (i.e., less than the average PMi0 risk estimates across cities), though not significant, predictors in
the Lippmann et al. (2006) analysis. Unlike the Franklin et al. analysis, arsenic (As) showed no
association in the Lippmann et al. (2006) analysis. The source of these differences may be due to the
difference in geographic coverage, PM size (PM2 5 may represent more secondary aerosols than PM10), or
December 2008
6-233
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
the difference in the analytical methods used in each study. Specifically, the analytical approach used by
Franklin et al. (2008) does have an advantage of delineating seasonal variations in PM components and
the associated potential seasonal mortality effects.
In light of the results presented in speciation studies it must be noted that second stage analyses that
use PM chemical species as effect modifiers have some limitations. Unlike analyses that directly examine
the associations between chemical species and mortality, if an effect modification is observed it may be
confounded if the variations of the mean levels of the chemical species examined are correlated with other
demographic factors that vary across cities. Thus, more concrete conclusions could be formulated if direct
associations are found between mortality and PM chemical components in time-series analyses.
Association between PM2.5 Chemical Components and Mortality in Six California
Counties
Ostro et al. (2007) examined associations between PM2 5 chemical components and mortality in six
California counties (Fresno, Kern, Riverside, Sacramento, San Diego, and Santa Clara), which had at least
180 days of speciation data for the years 2000 to 2003. The study examined all-cause, cardiovascular, and
respiratory mortality for individual lags of 19 specific PM2 5 chemical components. The second stage
random-effects model combined risk estimates at each lag across cities. The number of available days for
chemical species data ranged from 243 (San Diego County) to 395 (Sacramento County). The authors
found an association between mortality, especially cardiovascular mortality, and several chemical
components. For example, cardiovascular mortality was associated with EC, OC, nitrate, Fe, K, and Ti at
various lags.
Even though this was a multicity study, the relatively small number of available days and the
every-third-day (or every-sixth-day) sampling frequency for PM2 5 chemical species made it difficult to
interpret the results of the lag structure of associations observed for the chemical species. To evaluate the
impact of non-daily sampling frequency, Ostro et al. (2007) examined both the PM2 5 series that coincides
with the speciation sampling days (for the initial six counties [i.e., PM2 5c]) and PM2 5 data that was
available on all days for an extended set of counties (the initial six counties plus Contra Costa, Los
Angeles, and Orange Counties [i.e., PM2 5ext]). Figure 6-27 shows the association between all-cause
mortality and selected PM2 5 chemical species as well as for PM2 5c and PM2.5ext. Note the wide confidence
bands for the risk estimates for each PM2 5 chemical species and PM2 5c, apparently reflecting the low
statistical power of the data. The lag structure of associations is more clearly defined for PM2.5ext, and
appears to be different from that for PM2 5.
December 2008
6-234
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
300
DC
a
2.00
I 00
in
«
o
X.
U
=
4!
o
V
a.
-1.00
• •
• •
• it
• ~~
-200
0 12 3
0 1 2 3 0 12 3
EC 0C
0 12 3 0 12 3
NO,
so.
0 12 3
Cu
Species and lag day
0 12 3
K
0 1 2 3 0 1 2 3
Zn
PM.
2 5»*t
Source: Ostro et al. (2007)
Figure 6-27. Excess risk (CI) of total mortality per IQR of concentrations. Note: PM2.5 has the same
sampling days as chemical species. PM2.5 has all available PWh.sext data for nine
counties. * p < 0.10; ** p < 0.05
Ostro et al. (2008) used the speciation data from the six counties analyzed in their 2007 analysis,
described above, in an additional analysis to examine effect modification of cardiovascular mortality
effects, which showed the strongest association in the 2007 analysis, attributed to I'M; < and 13 chemical
components by socio-economic and demographic factors. The results of the analysis were combined using
random effects meta-analysis. The investigators tested statistical differences in risk estimates between
strata using a t-test, and reported that, for many of the PM2 5 chemical species; there were significantly
higher effect estimates among those with lower educational attainment and Hispanics. While these
patterns were apparent in their results table, interpretation of the results is not straightforward because the
table only presented the most significant (and positive) lags, and they were often different between the
strata (e.g., the most frequent significant lag for the Hispanic group was 1-dav, while it was 2- or 3-dav
for the White group). As the investigators pointed out, the every-third-day sampling frequency of the
speciation data also complicates the interpretation of the results for different lags.
Overall, the two studies by Ostro et al. (2007) were the first attempt to directly analyze associations
between the newly available chemical speciation data and mortality. While suggestive associations
between several chemical species and mortality were reported, a longer length of observations is needed
to more clearly determine the associations.
December 2008
6-235
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
6.5.2.6. Use of Source-Apportioned PM
Chemically speciated PM data allow for the source apportionment of PM. The idea of using
source-apportioned PM for health effects analyses is appealing because, if such source-apportionment
could be reliably conducted, it will allow for an evaluation of PM25 mass concentrations by source types.
However, the uncertainties associated with source-apportionment methods have not been well
characterized.
To address this issue, in 2003, several groups of EPA-funded researchers organized a workshop and
independently conducted source apportionment on two sets of data: Phoenix, AZ, and Washington, DC,
compared the results (Hopke et al., 2006), and then conducted time-series mortality regression analyses
using each group's source-apportioned data (Ito et al., 2006a; Mar et al., 2006; Thurston et al., 2005). The
various research groups generally identified the same major source types, each with similar elemental
compositions. Inter-group correlation analyses indicated that soil-, sulfate-, residual oil-, and
salt-associated mass concentrations were most unambiguously identified by various methods, whereas
vegetative burning and traffic were less consistent. Aggregate source-specific mortality relative risk (RR)
estimate confidence intervals overlapped each other, but the sulfate-related PM2.5 component was most
consistently significant across analyses in these cities.
The results from the source-apportionment workshop quantitatively characterized the uncertainties
associated with the factor analysis-based methods, but they also raised new issues. The mortality analyses
conducted in Phoenix, AZ, and Washington, DC, both found that different source-types showed varying
lag structure of associations with mortality. For example, Figure 6-28 shows cardiovascular mortality risk
estimates for three of the PM2 5 sources from the Phoenix, A Z, analysis (Mar et al., 2006). The strongest
associations for "traffic" PM2 5 was found for lag 1-day, while for "secondary sulfate" PM2 5, it was lag 0,
with a monotonic decline towards longer lags. It is conceivable that PM from different source types
produces different lagged effects, but it is also likely that different PM species have varying lagged
correlations with the covariates in the health effects regression models (e.g., temperature, day-of-week)
resulting in apparent differences in lagged associations with mortality. Thus, interpretation of these
source-apportioned PM health effect estimates remains challenging.
December 2008
6-236
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Secondary sulfate
Traffic
ABCDEFQH i ABCDEFGH ! ABODE F SHI
Source: Mar et al. (2006)
Figure 6-28. Relative risk and CI of cardiovascular mortality associated with estimated PM2.5
source contributions. Y-axis: relative risk per 5th-to-95th percentile increment of
estimated PM2.5 source contribution. X-axis: the alphabet denotes investigator/
method; lagged PM2.5 source contribution for lag 0 through 5 days, left to right, are
shown for each investigator/method.
6.5.2.7. Investigation of Concentration-Response Relationship
The results from large multicity studies reviewed in the 2004 PM AQCD suggested that strong
evidence did not exist for a clear threshold for PM mortality effects. However, as discussed in the 2004
PM AQCD, there are several challenges in determining and interpreting the shape of PM-mortality
concentration-response functions and the presence of a threshold, including: (1) limited range of available
concentration levels (i.e., sparse data at the low and high end); (2) heterogeneity of susceptibility in
at-risk populations; and (3) the influence of measurement error. Regardless of these limitations, studies
have continued to investigate the PM-mortality concentration-response relationship.
Daniels et al. (2004) evaluated three concentration-response models: (1) log-linear models (i.e., the
most commonly used approach, from which the majority of risk estimates are derived); (2) spline models
that allow data to fit possibly non-linear relationship; and (3) threshold models, using PMi0 data in 20
cities from the 1987-1994 NMMAPS data. They reported that the spline model, combined across the
cities, showed a linear relation without indicating a threshold for the relative risks of death for all-causes
and for cardiovascular-respiratory causes in relation to PM10, but "the other cause" deaths (i.e., all cause
minus cardiovascular-respiratory) showed an apparent threshold at around 50 (.ig/nr1 PM10, as shown in
Figure 6-29. For all-cause and cardio-respiratory deaths, based on the Akaike's Information Criterion
(AIC), a log-linear model without threshold was preferred to the threshold model and to the spline model.
The HEI review committee commented that interpretation of these results required caution, because
(1) the measurement error could obscure any threshold; (2) the city-specific concentration-response
December 2008
6-237
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
curves exhibited a variety of shapes; and (3) the use of AIC to choose among the models might not be
appropriate due to the fact it was not designed to assess scientific theories of etiology. Note, however, that
there has been no etiologically credible reason suggested thus far to choose one model over others for
aggregate outcomes. Thus, at least statistically, the result of Daniels et al. (2004) suggests that the
log-linear model is appropriate in describing the relationship between PMi0 and mortality.
CO
cc
>
o
cc
PM10 (pg/m3)
PM10 (pg/m3)
Other Cause Mortality
3 04 -
CVDRESP Mortality
Total Mortality
0 03 -
0.01 -
001
PM10 (pg/m3)
Source: Daniels et al. (2004)
Figure 6-29. Concentration-response curves (spline model) for all-cause, cardiovascular-
respiratory and other cause mortality from the 20 NMMAPS cities. Estimates are
posterior means under Bayesian random effects model. Solid line is mean lag,
triangles are lag 0 (current day), and squares are lag 1 (previous day).
The Schwartz (2004b) analysis of PMi0 and mortality in 14 U.S. cities, described in
Section 6.5.2.1., also examined the shape of the concentration-response relationship by including
indicator variables for days when concentrations were between 15 and 25 (.ig/nr1. between 25 and
34 (.ig/rn \ between 35 and 44 (.ig/nr1. and 45 (.ig/rn1 and above. In the model, days with concentrations
below 15 (.ig/rn1 served as the reference level. This model was fit using the single stage method,
combining strata across all cities in the case-crossover design. Figure 6-30 shows the resulting
relationship, which does not provide sufficient evidence to suggest that a threshold exists. The authors did
not examine city-to-city variation in the concentration-response relationship in this study.
December 2008
6-238
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
2.5
~
CD
"O
c
CD
1/5
a
a)
u
t=
c
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Total
Cardiovascular
Respiratory
0 50 100 160 200
PM10
Source: Samoli et al. (Samoli et al.)
Figure 6-31. Combined concentration-response curves (spline model) for all-cause,
cardiovascular, and respiratory mortality from the 22 APHEA cities.
Hie results from the three multicity studies discussed above support no-threshold log-linear
models, but issues such as possible influence of exposure error and heterogeneity of shapes across cities
remains to be resolved. Also, given the pattern of seasonal and regional differences in PM risk estimates
depicted in recent multicity study results (e.g., Peng et al., 2005), the very concept of the
concentration-response relationships estimated across cities and for all-year data may not be very
informative.
6.5.3. Summary of Causal Determinations by PM Metric
6.5.3.1. PM10
The epidemiologic evidence for the effect of short-term exposure to PM10 on mortality is
sufficient to conclude that a causal relationship is likely to exist at ambient concentrations. The
epidemiologic studies report consistent positive associations between short-term exposure to PM10 and
all-cause mortality. The multi-city studies evaluated reported effect estimates for all-cause mortality
ranging from 0.12% (Dominici et al., 2007b) to 0.81% (Schwartz, 2004b) per 10 |ig/m' increase in PM10.
Although respiratory- and cardiovascular-related mortality also show consislant positive effects, only a
few multi-city studies conducted cause -specific mortality analyses. The heterogeneity in the range of
these effects is most likely dependent on the lag, averaging time, number of cities and/or copollutants
included in regression models. Although respiratory and cardiovascular-related mortality also show
consistent, positive effects, only a few multicity studies conducted causes-specifici mortality analyses.
December 2008
6-240
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Risk estimates from single-city studies for all-cause and respiratory- and cardiovascular-related mortality
were similar to those from the multicity studies but consisted of smaller samples sizes which resulted in
wider confidence intervals and lower statistical power (Figure 6-20). An analysis of the lag structures
used in these studies found that the greatest effects were observed at lag 1 or lag 0-1, and the use of a
distributed lag model resulted in slightly larger (by -30%) estimates compared to single-day lags (Zeka et
al., 2005). The NMMAPS studies (Dominici et al., 2007b; Peng et al., 2005) confirmed the regional
heterogeneity in PMi0 risk estimates observed in previous analyses, with the greatest effects being
observed in the Eastern U.S. Seasonal patterns in risk estimates also were observed, but the results
differed between studies, with NMMAPS (Peng et al., 2005) showing the greatest effects during the
summer, while the 20 cities study (Zeka et al., 2005) observed the greatest effects during the transition
seasons, spring and fall. An analysis of potential confounders (i.e., temperature and copollutants) using
both time-series and case-crossover analyses found that neither is likely to account for differences in risk
estimates between PMio-mortality studies. However, one study (Burnett et al., 2004), observed a
reduction in the effect estimate upon the inclusion of N02 in the model.
6.5.3.2. PM2.5
The epidemiologic evidence on the effect of short-term exposure to PM2.5 on mortality is
sufficient to conclude that a causal relationship is likely to exist at ambient concentrations pm2 5
risk estimates were found to be consistently positive, and slightly larger than those reported for PMi0 for
all-cause, and respiratory- and cardiovascular-related mortality. The risk estimates for all-cause
(non-accidental) mortality ranged from 0.29% (Dominici et al., 2007b) to 1.21% (Franklin et al., 2007)
per 10 |ag/m3 increase in PM2 5. Cardiovascular-related mortality risk estimates were found to be similar
to those for all-cause mortality; whereas, the risk estimates for respiratory-related mortality were
consistently larger: 1.01% (Franklin et al., 2007) to 2.2% (Ostro et al., 2006) using the same lag and
averaging indices. Regional and seasonal patterns in PM2 5 risk estimates were observed with results
similar to those presented for PMi0, with the greatest effects occurring in the Eastern U.S. (Franklin et al.,
2007; Franklin et al., 2008) and during the spring (Franklin et al., 2007). Few studies conducted detailed
analyses of the potential confounding of risk estimates by gaseous pollutants, though Burnett et al. (2004)
analyzed gaseous pollutants and found mixed results, with possible confounding by N02. An examination
of effect modifiers (e.g., demographic and socioeconomic factors), specifically air conditioning use,
which is sometimes used as a surrogate for ventilation rate, has suggested that PM2 5 risk estimates
increase as the percent of the population with access to air conditioning decreases.
December 2008
6-241
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
6.5.3.3. PMio-2.5
The epidemiologic evidence on the effect of short-term exposure to PM10.2.5 on mortality is
suggestive of a causal relationship at ambient concentrations. Currently, a limited body of evidence
exists which has examined the potential association between short-term exposure to PM10_2.5 and
mortality. The majority of studies that examined PM10.2.5 reported mixed results in terms of the relative
impact of PM10-2.5 on mortality primarily due to the majority of studies being conducted in individual
cities and the city-to-city variation in the chemical composition of PMi0_2.5. However, one well conducted
multi-city Canadian study was identified that does provide evidence for an association between short-term
exposure to PMi0_2.5 and mortality, but these estimates along with those obtained from the other studies
may be confounded by gaseous copollutants or influenced by effect modifying conditions (e.g., season,
relative exposure error, dust storms). Overall, although more data is needed to adequately characterize the
chemical and biological components that may modify the potential toxicity of thoracic coarse particles,
the consistent association between short-term exposure to PMi0 (which includes PMi0_2.5) and mortality
provides some evidence for the presence of an association between PMi0.2.5 and mortality.
6.5.3.4. Ultra-fine particles (UFP: diameter: 0.01-0.1 \im)
The epidemiologic evidence on the effect of short-term exposure to UFP on mortality is
inadequate to infer a causal association at ambient concentrations. Only two new studies were
identified, which examined the association between short-term exposure to UFP and mortality.
Inconsistencies were observed in the lag structure of association reported by each study in terms of both
the lag day with the greatest association and the number of lag days considered in the study. The limited
number of studies and the discrepancy in results between studies further confirms the need for additional
data to examine the UFP-mortality relationship.
6.6. Attribution of Health Effects to Specific Constituents or
Sources
The chemical composition of PM may be a better predictor of health effects than particle size, but
despite the clear mechanistic plausibility of this hypothesis, only recently have researchers begun to
evaluate relationships between specific health effects and specific constituents or sources of ambient PM.
Prior to the 2004 PM AQCD, only a handful of epidemiologic studies attempted quantify these
December 2008
6-242
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
relationships. Approximately 30 new epidemiologic, human clinical, and toxicological studies have been
conducted to address this research question.
6.6.1. Evaluation Approach
The demands of conducting analyses that relate all PM constituents at once with an assortment of
health outcomes are very high for three reasons that are related to both the nature of PM data and
methodology. First, the number of PM constituents is large, and the correlations among them are
inherently high. Reducing this number has been accomplished in most of the recent studies through
various forms of factor analysis, which limits correlations by grouping the most highly correlated PM
constituents into less correlated groups. A subset of these studies identifies the resulting groups or factors
with named sources of PM. The methods for estimating source contributions to ambient PM are reviewed
in Section 3.5.4. Second, the number of potential health effects examined is also very high and includes
definitive outcomes (e.g., HAs) as well as physiologic alterations (e.g., increased inflammatory response).
Furthermore, examining the relationship between those two large sets of variables - PM constitutents or
sources on one hand, health variables on the other - involves analyzing the larger set of combinations
between them. Third, there are also multiple statistical methods and analytic approaches that can be used
to analyze the potential association.
A total of 34 studies are discussed below that attempted to link PM constituents or sources with
health outcomes, including four reviewed in the 2004 PM AQCD. All these studies included all measured
PM constituents in the analyses and most grouped PM constituents into specific PM sources. Studies that
used an a priori decision to consider only one or two constituents, or grouped PM constituents into only
two classes (e.g., soluble and insoluble), are not included. With the exception of the four studies reviewed
in the 2004 PM AQCD, and the toxicological in vitro studies, all remaining studies have been discussed
individually in Chapters 6 or 7 under the relevant health outcome. The studies evaluated include 14
epidemiologic studies, 15 toxicological studies, and five human clinical studies, and all used quantitative
methods to analyze possible associations between health effects and multiple PM constituents or sources.
Studies that presented PM composition and health data side by side without explicitly investigating
relationships or in which relationships were only discussed qualitatively, are excluded.
There are some differences in the type of PM constituent data used in epidemiologic, human
clinical and toxicological studies. In epidemiologic studies, PM speciation data is obtained from
atmospheric monitors; for human clinical and toxicological studies, the type of PM data varies with
experimental exposure technique. Thus, all 14 epidemiologic studies relied on monitor data, all of the
clinical studies and 11 of the toxicological studies used CAPs. The remaining four toxicology studies used
December 2008
6-243
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
PM samples collected on filters at various U.S. sites. Within the 15 toxicological studies evaluated, 11
conducted inhalation exposures, while suspensions were used in the other 5-4 consisted of in vitro
studies and 1 used in vivo intratracheal instillation.
Regardless of how PM was measured, all 34 studies provided data for between 7 and 20 PM
elements, with EC and OC, along with S04 and N03 also commonly measured. Most studies reduced the
number of PM variables by various factorization or source apportionment procedures, before using a
separate analysis to examine the association between the reduced PM dataset and health effects. In the
studies that reduced the number of PM variables, all human clinical studies and 8 of the toxicological
studies relied on PCA, while 2 more toxicological studies used CMB. Two toxicology studies used a
Partial Least Squares (PLS) procedure to simultaneously (1) reduce the PM data, and (2) examine the
association between PM and health effects, rather than decoupling the two parts of the analysis. One of
these two studies used both CMB and PLS. One toxicology study used a Structural Equation Model
(SEM) to simultaneously reduce the PM data and examine the association between PM and health effects.
Finally, three toxicological and two human clinical studies did not apply any kind of grouping to the
speciation data. All but three epidemiologic studies reduced the number of PM variables, using either
PCA, or other established source apportionment methods such as CMB, PMF, or UNMIX (Table 6-16).
Factorization and apportionment methods were reviewed in Section 3.5.4.
Of particular interest are five epidemiologic studies that compared source apportionment methods
and the associated results. (Ito et al., 2006a; Mar et al., 2006; Thurston et al., 2005), compared PCA,
PMF, and UNMIX (Hopke et al., 2006) independently applied by separate research groups. Schreuder et
al. (2006) compared UPM and two versions of UNMIX and Sarnat et al. (2008) compared CMB, PMF,
and literature-based a priori groupings. In all five, results based on the different methods were generally
in close agreement. When applied to monitoring data, the explicit aim of many of these grouping or
factorization methods is to apportion PM species to their most likely sources. All but three epidemiology
studies labeled the groupings according to their presumed common source. However, only four toxicology
studies and two clinical studies explicitly named PM sources corresponding to factors. Eight out of the 11
toxicological studies that employed CAPs and grouped PM species, but did not name sources, were
conducted in Boston by the same research group.
One important difficulty in interpreting these 34 studies as a group is that few, if any of the results
of the various grouping procedures are easily comparable, due to both differences in PM constituents that
comprise the factors identified in separate studies, and the subjectivity involved in labeling those factors.
There are no well-established objective methods for operating the various forms of factor analysis and
source apportionment used in these studies, leaving much of the model operation and factor assignment
open to judgment by the individual investigator. For example, it is not possible to easily compare the
December 2008
6-244
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Al/Si factor isolated in one study, with the Al/Ca/Fe/Si factor from another study, and the "Resuspended
Soil" factor from a third study.
After factorization or apportionment of the PM data, various methods were used for analyzing the
potential association between PM constituents or sources and health effects. Except for the three studies
that used PLS or SEM, and thus did not decouple the two phases of analysis, human clinical and
toxicological studies all used univariate mixed model regression for every PM factor or source. One of
these toxicological studies was of subchronic duration, and repeated samples were collected over an
extended period, thus supporting the addition of regressional effect of time. Six toxicological studies
followed the univariate step with multivariate regression for all factors. Epidemiologic studies generally
related sources with health outcomes through various forms of Poisson regression, and four used GAMs.
One epidemiologic study used linear regression, and one long-term exposure study used time-to-event
data analysis (survival analysis) methodology.
6.6.2. Findings
6.6.2.1. Results from the 2004 PM AQCD
Four epidemiologic studies were evaluated in the 2004 PM AQCD that examined the association
between PM constituents or sources and specific health effects. Of these studies, one study associated
daily mortality with mobile sources in Knoxville, TN and St. Louis, MO and coal in Boston, MA; while
the crustal factor was not found to be significant for any of the six cities studied (Laden et al., 2000).
Another study demonstrated an association between regional sulfate and total mortality at lag 0 in
Phoenix and regional sulfate, motor vehicles, and vegetative burning with cardiovascular mortality at lags
of 0, 1, and 3, respectively (Mar et al., 2000; Mar et al., 2003). Negative associations were observed
between total mortality and regional sulfate at lag 3, along with local S02 and soil factors (Mar et al.,
2000; Mar et al., 2003). Finally, Tsai et al. (2000) identified significant associations between
PMi5-derived industrial sources and total daily deaths in Newark and Camden, NJ; sulfate was also linked
to cardiopulmonary deaths in both locations. Total mortality and cardiopulmonary deaths were also
significantly associated with oil burning sources in Camden (2000).
6.6.2.2. Epidemiologic Studies
For the comparative study described in Hopke et al. (2006), the results of which are reported in
Thurston et al. (2005), Ito et al. (2006a), and Mar et al. (2006), total (nonaccidental) mortality was
associated with secondary sulfate in both Phoenix and Washington D.C., although lag times differed (0
December 2008
6-245
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
and 3, respectively) (Ito et al., 2006a; Mar et al., 2006). Primary coal was also associated with total
mortality in Washington D.C. (Ito et al., 2006a) and copper smelter, traffic, sea salt, and biomass/wood
burning were associated with cardiovascular mortality in Phoenix at various lag times (Mar et al., 2006).
In an additional source apportionment analysis, Sarnat et al. (2008) found in Atlanta that
cardiovascular disease-related ED visits were associated with same-day mobile sources (gasoline and
diesel) and biomass combustion (primarily prescribed forest burning and residential wood combustion),
whereas secondary sulfate was associated with respiratory disease ED visits (Sarnat et al., 2008). Sarnat et
al. (2008) also found that the power plant source identified by the CMB approach was negatively
associated with respiratory ED visits while no association was found for soil and secondary
nitrates/ammonium nitrate.
Other epidemiological studies (i.e., panel studies) have examined the association between PM
sources and physiologic alterations in cardiovascular function. Lanki et al. (2006a) reported associations
between local traffic and ST-segment depression in elderly adults in a study conducted in 3 European
cities (i.e., Helsinki, Amsterdam, and Erfurt) (Lanki et al., 2006a). Yue et al. (2007) found that adult males
with coronary artery disease demonstrated changes in repolarization parameters associated with
traffic-related PM, with increased vWF linked to traffic and combustion-generated particles (Yue et al.,
2007). In addition, elevated CRP levels were associated with all sources (soil, local traffic, secondary
aerosols from local fuel combustion, diesel, and secondary aerosols from multiple sources). Reidiker et al.
(2004b), in a study of young highway patrol officers, found that the most significant effects (HRV,
supraventricular ectopic beats, hematological markers, vWF) were associated with a "speed-change"
factor (Riediker et al., 2004b). In addition, Riediker et al. (2004b) observed an association between crustal
factor and cardiovascular effects, but no association with steel wear and gasoline.
The only epidemiologic study that evaluated respiratory ED visits was conducted in Spokane, WA
and used tracers as indicators of PM sources (Schreuder et al., 2006). In this study, only vegetative
burning (total carbon) was associated with increased respiratory ED visits while motor vehicles (Zn) and
soil (Si) were not associated with any health outcomes.
There were four epidemiologic studies that did not group PM constituents and differed in study
design. In a study of long-term exposure to PM2 5, total mortality was associated with EC in a cohort of
U.S. military veterans (Lipfert et al., 2006b). Long-term exposure to Ni and V on daily risk of total
mortality was found in residents of New York (Lippmann et al., 2006). Short-term exposure to PM2.5
constituents was associated with total mortality for EC, OC, S04, Ca, Fe, K, Mn, Pb, S, Si, Ti, and Zn in
winter in a cohort of six California counties (Ostro et al., 2007). Similarly, cardiovascular mortality was
associated with EC, N03, S04, Fe, K, S, Ti, and Zn in susceptible individuals living in six California
December 2008
6-246
DRAFT-DO NOT QUOTE OR CITE
-------
1 counties (Ostro et al., 2008), plus OC, Mn, and V for Ostro et al. (Ostro et al., 2007). For the latter two
2 studies, lag times were 2 or 3 days.
Table 6-16. Epidemiologic studies of PM sources, factors, or individual components.
Reference:
Lippmann et al.
(2000)
Location: rural
location upwind
from NYC
Reference: Mar
et al. (2000)
Location: one
monitor in
Phoenix, AZ
Reference: Tsai
et al. (2000)
Location: 3 NJ
sites for 2
summers
(ATEOS study)
Subjects: mice
(ApoE)
Exposure: CAPs
(avg. mass
concentration 85.6
Mg/m3)
n: 12 ApoE mice
(6/g roup)
Number of
constituents
considered for
grouping: NR
Grouping method: Groups/Factors/
no grouping was Sources: NR
performed
# of groups: NR
PM variables used:
Mass contribution of
every constituent in
CAPs portion of
study, contribution of
16 constituents in epi
portion
Results: Lag for HR elevations on 14 days with wind from NW was same day. Lag for SDNN reduction on 14 days with wind from NW was 0,
1 and 2.
GAM analysis: Beta coefficient significant for Ni and HR (but not V, Cr, or Fe). Beta coefficient significant for Ni and log SDNN (but not V, Cr,
or Fe).
Epi results: The strongest associations were between PM10 mortality risk and Ni and V.
Reference:
Laden et al.
(2000)
Location:
Monitors in 6
Eastern US cities
(Harvard Six
Cities Study)
Subjects: NR
Exposure: NR
n: NR
Number of
constituents
considered for
grouping: 15
elements
Grouping method:
PCA
# of groups: 8
Groups/ Factors/
Sources:
Soil/Crustal (PM
fine), Mobile vehicle
exhaust (PM fine),
Coal (PM fine), Fuel
oil; Metals, Salt'
Manganese,
Residual
PM variables used:
Tracers: Si, V, CI,
Pb.Se
Results: Lag 0-1 average for all results. Over all 6 cities, mobile source factor (Pb) had greatest association with daily mortality (3.4%) with
10 ug/m3 increase. The greatest effects for mortality due to mobile sources were observed in Madison (Portage), Knoxville
(Kingston-Harriman), and St. Louis, althugh the Madison results were not statistically significant. The coal source factor was only significant
in Boston (Watertown) - 2.8% increase in mortality and the overall percent increase was also significant (1.1%). Deaths from pneumonia
attributable to coal combustion sources was 7.9% (CI, 3.1-12.7%) and statistically significant. The crustal factor was not associated with
mortality in any city, although this factor was not a significant predictor in the regression model for Boston (Watertown) due to its low
contribution to PM2.5 mass. For specific elements included simultaneously, sulfur, Pb, and Ni were significantly associated with overall
mortality (3.0,1.6,1.5%, respectively). Boston had the greatest percent increase in mortality for sulfur (7.9%), Knoxville for Pb (15.0%), and
Steubenville for Ni (8.2%), although the CIs are all quite large.
Reanalysis results (Schwartz et al., 2003): Effects changed slightly. New percent increases in mortality for combined cities are 3.5 and 0.79
for traffic and coal, respectively. The coal factor in Boston decreased to 2.1% increased mortality. A residual oil factor in Boston and
Steubenville resulted in at 22.9% increase in daily deaths (but was not significant in the original analysis).
Subjects: elderly
only
Exposure: NR
n: NR
Number of
constituents
considered for
grouping: 10
elements, OC, EC,
CO, N02; S02
Grouping method:
unspecified type of
factor analysis
# of groups: 3 or 5
Groups/ Factors/
Sources: motor
exhaust/road dust,
soil, vegetative
burning, local SO2,
regional SO4
PM variables used:
first used individual
constituents: S, Zn,
Pb, K, OC, EC.TC
(AL+Si+Ca+Fe+Ti),
then factor scores
Results: Cardiovascular mortality associated with PM2.5 mass on lag 1 and 4 (6 and 4%, respectively). EC and TC associated with CV
mortality for lag 1 (RR = 1.05); OC was weakly associated with CV mortality for lags 1 and 3. For total mortality, regional suflate was
positively assocated at lag 0, but negatively associated at lag 3. The local SO2 and the soil factors were nagatively associated with total
mortality. For CV mortality, secondary sulfate was positively associated at lag 0, motor vehicle at lag 1, and vegetative buring at lag 3.
Reanalysis results (Mar et al., 2003): Similar associations were observed.
Subjects: NR
Exposure: NR
N: NR
Number of
constituents
considered for
grouping: 8 metals,
IPM, FPM, S04, CX,
DCM, ACE, CO
Grouping method:
unspecified type of
factor analysis
# of groups: 5
Groups/ Factors/
Sources: oil burning,
motor emissions,
resuspended dust,
secondary aerosol,
industrial sources
PM variables used:
Used individual
constituents, then
factor scores, then
tracers
Results: RR associated with 10 ug/m3 increases: Newark -1.03 for industrial and total daily deaths; 1.02 for sulfate and total daily deaths;
1.04 for sulfate and cardiopulmonary deaths. Camden -1.11 for oil burning sources and total daily deaths; 1.10 industrial and total daily
deaths; 1.12 for oil buring sources and cardiopulmonary daily deaths; 1.02 for sulfate and cardiopulmonary daily deaths
December 2008
6-247
DRAFT-DO NOT QUOTE OR CITE
-------
Reference:
Riediker et al.
(2004)
Location: Inside
9 state police
patrol cars
Reference:
Lanki et al.
(2006a)
Location:
Monitors in
Helsinki, Finland,
Amsterdam,The
Netherlands and
Erfurt, Germany
Reference:
Mar et al.
(2006)
Location:
Phoenix, AZ
Reference:
Schreuder et al.
(2006)
Location: One
monitor in
Spokane, WA
for 7 years
Subjects: healthy
male young police
officers
Exposure: 9, on 4
consecutive days
n: NR
Number of
constituents
considered for
grouping: 10
elements; 3 gaseous
pollutants; 2 physical
variables
Grouping method:
PCA
# of groups: 4 when
13+2 constituents
included; 3 when
only 9
"PM-associated"
constituents included
Groups/ Factors/
Sources: soil;
automotive steel
wear; gasoline
combustion;
speed-changing
traffic
PM variables used:
Mass contribution or
score (?) of sources
Results: Using two different factor analysis models found most significant effects (MCL, SDNN, PNN50, supraventricular ectopic beats, %
neutrophils, % lymphocytes, MCV, von Willebrand Factor, and protein C) were for "speed-change factor" (i.e., copper, sulfur, aldehydes).
Some associations observed for "crustal" and none for "steel wear" and "gasoline."
Reference: Ito et
al. (2006a)
Location:
Washington, DC
Subjects: NR
Exposure: NR
n: NR
Number of
constituents
considered for
grouping: NR
Grouping method:
comparison of: PMF;
(absolute) PCA;
UNMIX
# of groups: 6-10
Groups/ Factors/
Sources: Different
research groups
gave different names
to sources (see
Hopke et al, table 2)
Sources for which
association with
health was
analyzed: Soil,
traffic, Secondary
S04, NOs (Wash DC
only), residual oil
(Wash DC only),
Wood smoke/
biomass combustion,
Sea salt, incinerator
(Wash DC only),
primary coal (Wash
DC only), Cu smelter
(Phoenix only)
PM variables used:
mass contribution of
sources
Results: Overall, PM2.5 effects observed at lag 3. Lag structure of association varied across source types, but consistent across investigators
for total (non-accidental mortality): soil factor - mostly positive at various lags (not significant); secondary sulfate - strongest association at lag
3; nitrate - mostly negative except at lag 3; residual oil - strongest association at lag 2 (not significant); wood-burning - increasing association
as lag increases (not significant); incinerator - significant negative associations at lag 0; primary coal - significant assocation at lag 3.
Subjects: NR
Exposure: NR
n: NR
Number of
constituents
considered for
grouping: 13
elements
Grouping method: Groups/ Factors/ PM variables used:
Absolute PCAx
# of groups: 5
Sources: crustal;
long range
transported; oil
combustion; soil;
traffic
Tracers: Si (crustal);
S (long-range
transport); Ni (oil
combustion); CI
(salt); ABS (local
traffic).
Results: Highest observed effects were for crustal sources and salt at lag 3 (when analyzing sources), but not consistent or significant. In
multipollutant models only ABS associated with ST-segment depression, but wide CIs. When examining indicator elements of a source, local
traffic found to be the most toxic, but when examined per IQR long-range transport and traffic had similar effects.
Subjects: NR
Exposure: NR
n: NR
Number of
constituents
considered for
grouping:
Unknown
Grouping method: Groups/Factors/ Sources for which PM variables
comparison of:
PMF (absolute);
PCA; UNMIX
# of groups: 6-10
Sources: Different
labs gave different
names to sources
(see Hopke et al,
table 2)
association with
health was ana-
lyzed: Soil, Traffic,
secondary SO4,
NOs, (Wash DC
only), residual oil
(Wash DC only),
woodsmoke/
biomass combus-
tion, sea salt,
incinerator (Wash
DC only); primary
coal (Wash DC
only); Cu smelter
(Phoenix only)
used: mass
contribution of
Results: Using daily PM2.5 data found the following associations with cardiovascular mortality: Secondary sulfate - greatest effect observed
for all sources and at lag 0; traffic - associated at lag 1; copper smelter associated at lag 0; sea salt - had the greatest statistical significance
and observed at lag 5; biomass/wood burning - less consistent lag structure but greatest association at lag 3; soil - did not show an
association or consistent lag structure. For total (non-accidental) mortality associations were weaker and consistently observed for only:
copper smelter - lag 0; sea salt - lag 5.
Subjects: NR
Exposure: NR
n: NR
Number of
constituents
considered for
grouping: 11
elements, TC, NO3
Grouping method: Groups/Factors/ PM variables
Comparison of:
PMF, UNMIX,
Multilinear Engine
# of groups: 8
Sources:
Vegetative burning;
Ao.rirh
Vehicle; S04; NOs;
Soil; Cu-rich;
Marine
used: Tracers: TC
(vegetative
burning); As
(As-rich); Zn
(vehicle); Si (soil)
Results: Si, As, and Zn were not associated with any health outcomes; while an IQR increase in TC (vegetative burning) was associated
with a 2% increase in respiratory ED visits.
December 2008
6-248
DRAFT-DO NOT QUOTE OR CITE
-------
Reference:
Yue et al.
(2007)
Location: one
monitor in
German city,
30.000 samples
Subjects: adult n: 56, data Number of
males collected 12 times constituents
Exposure'CAD over 6 month for considered for
every subject, but grouping:
extended period Apportionment
missing PM data based on particle
size distribution.
Grouping method: Groups/Factors/
PMF Sources: airborne
# of groups: 5 soil, local traffic,
loacal fuel
combustion, remote
traffic (diesel),
secondary aerosols
PM variables
used: Mass
contribution or
score (?) of
sources
Results: Overall, repolarization parameters influenced by traffic-related particles; vWF increased in response to traffic-related particles and
combustion-generated aerosols. All source factors contributed to increasing CRP levels.
Reference:
Sarnat et al.
(2008)
Location: one
monitor in
Atlanta. GA for
2 yrs
Subjects: NR n: NR Number of
Exposure: NR constituents
considered for
grouping: NR
Grouping method: Groups/Factors/
Comparison of: Sources: gasoline,
PMF, CMB-LGO, diesel, wood
SOP ("a priori") smoke/biomass
# of groups: 9,11 burning, soil
(6 of them common secondary SOV
between methods) ammonium sulfate,
secondary nitrate/
ammonium nitrate,
metal processing,
railroad, bus and
highway, cement
kiln, power plants,
other OC,
ammonium
bisulfate
PM variables
used: Mass
contribution or
score (?) of
sources and tracers
Results: Sulfate secondary associated with 1.2 - 2.0% increase in RD visits, significant negative association RD visits and primary
emissions from coal-fired power plants. CVD significantly associated with other OC (1.014), biomass (1.033), diesel and gas for CMB-LGO.
For PMF and CVD visits: diesel (1.025), gas, wood smoke, metal processing (1.013). Year-long associations: PMF diesel, EC, CMB-LGO
gas, Zn and biomass combustion sources (CMB-LGO biomass burning, PMF wood smoke, and K). Diesel and gas sources association with
RD in the warm season (1.2-2.1% per IQR).
Reference:
Lipfert et al.
(2006a)
Location:
Many US
locations
(STN), large
cohort of US
Veterans
Subjects n/a n: n/a Number of
Exposure: n/a constituents
considered for
grouping: 15
elements, EC, OC,
S04, NOs
Grouping method: Groups/Factors/
no grouping was Sources: mass
performed contribution of
# of groups: every constituent,
and other
pollution-related
variables
PM variables
used: Time to
event analysis
(semi parametric
proportional
hazards
model = Cox
regression), single
and multi pollutants
models, some
including several
constituents of PM,
non-PM pollutants,
and other
pollution-related
variables.
Results: All showed significant associations with mortality with traffic density and EC showing the greatest effects.
Reference:
Ostro et al.
(2007)
Location:
monitors in 6
CA counties,
some with 2
Subjects n/a n: n/a Number of
Exposure: n/a constituents
considered for
grouping: 15
elements, EC, OC;
NOs; S04, PM2.5
mass
Grouping method: Groups/Factors/
no grouping was Sources:
performed
# of groups: n/a
PM variables
used: mass
contribution of
every constituent
monitors, for 4
years
Results: Effects were greater during the winter months. In the all year analysis, at 3-day lag associations observed for EC, OC, NO3 and
Zn. During winter months (Oct -March) effects observed for most species for both all-cause and cardiovascular mortality at lag 3 (EC, OC,
SO4, Ca, Fe, K, Mn, Pb, S, Si, Ti, Zn) and (OC, NO3, SO4, Fe, Mn, S, V, Zn), respectively.
Reference:
Ostro et al.
(2008)
Location:
monitors in 6
CA counties,
some with 2
Subjects n/a n: n/a Number of
Exposure: n/a constituents
considered for
grouping: 9
elements,EC, OC,
PM25 mass, SO4,
NOs
Grouping method: Groups/Factors/
no grouping was Sources:
performed
# of groups:
PM variables
used: mass
contribution of
every constituent
monitors/4
years
Results: The following associations were observed with cardiovascular mortality: PM2.5 (lag 3); EC (lag 2); NO3 (lag 3); SO4 (lag 3); Fe (lag
2); K (lag 2); S (lag 3); Ti (lag 2); Zn (lag 3).
December 2008
6-249
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
6.6.2.3. Human Clinical Studies
Three human clinical studies employed PCA, although only one linked groupings of PM
constituents to the measured physiologic parameters (Table 6-17). Huang et al. (2003c) demonstrated
associations between increased fibrinogen and Cu/Zn/V and increased BALF neutrophils and Fe/Se/S04
in young, healthy adults exposed to RTP, NC CAPs; however, only water-soluble constituents were
analyzed. In the other study that examined physiologic cardiovascular effects, Fe and EC were associated
with changes in ST-segment, while sulfate was associated with decreased systolic BP in asthmatic and
healthy human volunteers exposed to Los Angeles CAPs (2003a). In Gong et al. (2003a) the majority of
the PM was in the thoracic coarse fraction. In the other study that used Los Angeles CAPs, the only
observed association was between sulfate content and decreased lung function (FEVi and FVC) in elderly
volunteers with and without COPD (2005). Two additional human clinical studies that did not perform
grouping and employed Toronto CAPs plus ozone demonstrated increased diastolic BP and increased
brachial artery vasoconstriction associated with carbon content (Urch et al., 2004; Urch et al., 2005).
Table 6-17. Human clinical studies of PM sources, factors, or individual components.
Study:
Gong et al.
(2003a)
Location:
Los
Angeles,
CA
Study:
Gong et al.
(2005)
Location:
Los
Angeles,
CA
Reference:
Huang et
al. (2003c)
Location:
Chapel Hill,
NC
Reference:
Urch et al.
(2004)
Location:
Toronto,
Canada
Reference:
Urch et al.
(2004)
Location:
Toronto,
Canada
Subjects: Adult 18-45,
healthy vs. asthmatic
Exposure: CAPs healthy
and asthmatic exposed at
different times
N: 12 healthy, 12 asthmatic Grouping method: PCA
Constituents considered # of groups: 4 (note: OC
for grouping: 7 elements, data was unavailable)
EC, NOs, S04
Factors/Source: crustal (Al PM variables used: Total
Si CA K Fe) mass
S (2 metrics of S04 + Tracers: S04, EC, Fe
elemental S)
Total Mass+N03, EC
Results: Fe and EC associated with a decrease in ST-segment voltage 2 days post-exposure. EC associated with an increase in ST-segment
voltage immediately following exposure. Sulfate content associated with a decrease in systolic BP 4 h post-exposure.
Subjects: elderly, COPD
vs. healthy/ CAPs
Exposure: NO2 (full
factorial)
N: 6 healthy, 18 COPD
Constituents considered
for grouping: 7 elements
+ EC
Grouping method: PCA
# of groups: 3 (note: OC
was unavailable)
Factors/Source: crustal (Al PM variables used: Total
Si CA K Fe) mass, Tracers:, SO4, Si, Fe,
S (= S04) EC
Na
Results: Mass concentration of CAPs not observed to significantly affect lung function. However, sulfate content was associated with a decrease
lung function (FEV1 and FVC), which was enhanced by co-exposure to NO2.
Subjects: healthy adults
Exposure: CAPs
N: 35 male; 2 female
Constituents considered
for grouping: 8 elements
and SO4
Grouping method: PCA
# of groups: 2
Factors/Source:
Fe/SOVSe/V/Zn/Cu
PM variables used:
factor scores, then mass
contribution of all 9
constituents
Results: Associations observed between sulfate, zinc, and selenium content and increases in BAL neutrophils. Increases in fibrinogen associated
with copper, zinc, and vanadium content.
Subjects: healthy adults
19-50 yrs/CAPs
Exposure: O3
N: 23
Constituents considered
for grouping: unknown
Grouping method: no
grouping was performed
# of groups: NR
Factors/Source:
NR
PM variables used:
every constituent in uni-
variate analysis, then OC
and SO4 in multivariate
analysis
Results: CAPs-induced increase in diastolic BP significantly associated with carbon content of the particles.
Subjects: healthy
adults/CAPs
Exposure: O3
N: 24 Grouping method: no
Constituents considered 9rouPln9 was Performed
for grouping: 14 elements, # of groups: NR
EC, OC
Factors/Source:
NR
PM variables used: every
constituent in univariate
analysis, then OC and SO4
in multivariate analysis
Results: Both organic and EC content of CAPs associated with an increase in brachial artery vasoconstriction.
December 2008
6-250
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
6.6.2.4. Toxicological Studies
The single toxicological in vivo study that characterized PM sources corresponding to identified
sources was conducted in Tuxedo, NY, over a 5-month period (Table 6-18). This study reported that all
sources (regional sulfate, resuspended soil, residual oil, traffic and other unknown sources) were linked to
HR or HRV changes in mice at one time or another during or after daily exposure (Lippmann et al.,
2005a). In a similar in vitro study using CAPs from the same location for an in vitro exposure, NF-kB in
BEAS-2B cells were correlated with the oil combustion factor (Maciejczyk and Chen, 2005). The other in
vitro toxicological study (Duvall et al.) that named sources employed samples from five U.S. cities and
found a high correlation between increased IL-8 release in primary human airway epithelium cells and
coal combustion (R2=0.79), secondary nitrate (R2=0.63), and mobile sources (R2=0.39). In addition, soil
(R2=0.48), residual oil combustion (R2=0.38), and wood combustion (R2=0.33) were associated with
COX-2 effects; whereas, secondary sulfate (R2=0.51) was correlated with HO-1. Wood combustion and
soil were negatively correlatted with HO-1.
There were six toxicological studies that employed Boston CAPs and identified at least four
groupings of PM constituents (V/Ni, S, Al/Si, and Br/Pb), but only partially and tentatively named
sources (Batalha et al., 2002; Clarke et al., 2000; Godleski et al., 2002; Nikolov et al., 2008; Saldiva et al.,
2002; Wellenius et al., 2003). When examining cardiovascular effects these studies found that Si was
associated with changes in the ST-segment of dogs (Wellenius et al., 2003) and decreased lumen/wall
ratio in rat pulmonary arteries (Batalha et al., 2002) in multivariate analyses. In addition, blood
hematological results were associated with V/Ni, Al/Si, Na/Cl, and S in dogs (Clarke et al., 2000). An
examination of respiratory effects found that V/Ni and Br/Pb were associated with increased
inflammation in BALF for only the 3rd day of exposure (Clarke et al., 2000). Decreased respiratory rate
and increased airway irritation (Penh) in dogs were associated with road dust (Al) and motor vehicles
(OC), respectively (Nikolov et al., 2008). Individual PM2 5 constituents associated with elevated
neutrophils in BALF were Br, EC, OC, Pb, and sulfate (Godleski et al., 2002), which is consistent with
the findings (Br, EC, OC, Pb, V, and CI) of Saldiva et al. (Saldiva et al., 2002). Univariate regression of
two Boston CAPs studies that did not group PM constituents demonstrated that lung oxidative stress was
correlated with Mn, Zn, Fe, Cu, and Ca (Gurgueira et al., 2002) and Al, Si, Fe, K, Pb, and Cu (Rhoden et
al.) in rats.
The two toxicological studies that used PLS methodologies identified PM2 5 constituents linked to
respiratory parameters. Seagrave et al. (2006) demonstrated associations between cytotoxic responses and
gasoline plus nitrates (OC, Pb, hopanes/steranes, nitrate, and As) along with inflammatory responses and
gasoline plus diesel (including major metal oxides) in rats exposed via intratracheal instillation. In the
other study, Veranth et al. (2006) collected loose surface soil from 28 sites in the Western U.S. and
December 2008
6-251
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
exposed BEAS-2B cells to PM2 5. In the multivariate redundancy analysis, OC1, OC3, OC2, EC2, Br,
EC1, and Ni correlated with IL-8 release, decreased IL-6 release, and decreased viability at low and high
doses (10 and 80 |ig/cm2. respectively).
One in vitro toxicological study that employed Chapel Hill PM used PCA but did not name specific
PM sources (Becker et al., 2005b). In this study, the release of IL-6 from human alveolar macrophages
and IL-8 from normal human bronchial epithelial cells was associated with a PMi0 factor comprised of
Cr, Al, Si, Ti, Fe, and Cu. No statistically significant effects were observed for a second PM factor (Zn,
As, V, Ni, Pb, and Se).
Those toxicological studies that did not apply groupings to the PM speciation data demonstrated a
variety of results. As reported above, two Boston CAPs studies demonstrated lung oxidative stress
correlated with Mn, Zn, Fe, Cu, and Ca (Gurgueira et al., 2002) and Al, Si, Fe, K, Pb, and Cu (Rhoden et
al., 2004) in rats. The remaining toxicological study that did not use PM species groupings reported a
correlation between Zn and plasma fibrinogen in SH rats when constituents were normalized per unit
mass of CAPs (Kodavanti et al., 2005).
Table 6-18. Toxicological studies of PM sources, factors, or individual constituents
Reference:
Becker et al.
(2005a)
Location:
Chapel Hill,
NC;
repeated
sampling for
1 year
Subjects: normal human
bronchial epithelial and
human AM
Exposure: (2-3x105
cells/ml; 11 or 50 ug/ml)
n: not provided Grouping method: PCA
Constituents considered # of groups: 2
for grouping: 12 elements
Groups/ Factors/ PM variables used: not
Sources: Cr/AI/Si/Ti/Fe/Cu provided
("crustal"), Zn/AsA//Ni/Pb/
Se
Results: Cr/AI/Si/Ti/Fe/Cu associated with IL-8 release in normal human bronchial epithelial cells and IL-6 release in AM. Zn/As/V/Ni/Pb/Se not
associated with any endpoints. Stepwise linear regression with individual constituents Fe and Si assoicated with IL-6 release in AM. Cr associated
with IL-8 release in NHBE cells.
Reference:
Batalha et al.
(2002)
Location:
Boston, MA
Reference:
Clarke et al.
(2000)
Location:
Boston, MA
Subjects: rats
Exposure: CAPs (3-day
mean CAPs concentration
range: 126.1-481.0 ug/m3)
CAPs (3-day mean CAPs
concentration range:
126.1-481.0 ug/m3)
n: 7-10 rats x 2 levels
CAPs x 2 levels SO2 x 6
runs in different seasons
Constituents considered
for grouping: 20 elements;
OC; EC
Grouping method:
Previous study in same city
(Clarke et al.), and PCA of
this experiment's data
# of groups: 6 independent
variables in univariate
regression (4
elements,EC,OC)4 in the
multivariate step (all of
them tracers)
The 4 tracer elements
corresponded to the 4
groupings identified by
PCA)
Groups/ Factors/
Sources: V/Ni, S, Al/Si,
Br/Pb
PM variables used:
Tracers: Si, SO4, V, Pb.
Other: EC, OC
Results: Univariate analyses for first day not significant for L/W ratio. Univariate analyses for second and third day and second+third day mean
were similar. Presented second+third day mean regression data. CAPs mass, Si, Pb, SO4, EC, OC significant for decreased LA/V ratio in
normal+CB rats exposed to CAPs. Si, SO4 significant for decreased L/W ratio in normal rats. Si, OC significant for decreased L/W ratio in CB rats.
Multivariate analysis using normal+CB rats for Si, SO4, V, Pb - only Si remained significant with decreased L/W ratio.
Subjects: dogs
Exposure: CAPs (average
for all studies, paired:
203.4, crossover: 360.8
ug/m3) repeated exposure
with several weeks in
between
n: 10 dogs, 20 paired
exposures, 24 crossover
Constituents considered
for grouping: 19 elements,
black C
Grouping method: PCA
# of groups: 4 for
exposure in paired runs
6 for exposure in crossover
runs
Groups/ Factors/
Sources: V/Ni, S, Al/Si,
Br/Pb, S, Na/CI
Cr
PM variables used: all
elements, then factor
scores
December 2008
6-252
DRAFT-DO NOT QUOTE OR CITE
-------
Results: No significant differences between baseline, sham, or CAPs group for BAL cell differential percentages.Total BAL protein increased with
CAPs compared to sham. No significant hematological effects with CAPs exposure. Mixed linear regression analyses (statistics not provided): Al
and Ti (3-day ave. concentrations) associated with dose-dependent decreases in BAL AM and increases in BAL PMN percentages. Sulfate
associated with increased WBC. BC, Al, Mn, Si, Zn, Ti, V, Fe, Ni associated with increased blood PMN. Na associated with increased blood
lymphocytes. Al, Mn, Si associated with decreased blood lymphocytes. CAPs mass and BC associated with decreased blood eosinophils. CAPs
mass associated with decreased platelet count. Regression using results of factor analysis: None for 3-day ave. concentration for BAL
parameters. V/Ni for increased AM percentage and Br/Pb for increased PMN percentage for 3rd-day only concentration. V/Ni and Al/Si for
increased blood PMN percentage and decreased blood lymphocyte percentage. Al/Si also for increased WBC counts. Na/CI for increased blood
lympyocyte percentage. S for decreased RBC and hemoglobin.
Reference:
Godleski et
al. (2002)
Location:
Boston, MA
Subjects: rats
Exposure: CAPs (3-day
mean CAPs concentration
range: 126.1-481.0 ug/m3)
n: 7-10 rats x 2 levels
CAPs x 2 levels SO2 x 6
runs in different seasons
Constituents considered
for grouping: 20 elements,
OC, EC
Grouping method:
Previous study in same city
(Clarke et al.), and PCAof
this experiment's data
# of groups: 6 independent
variables in univariate
regression (4
elements,EC,aOC)
Groups/ Factors/
Sources: V/Ni, S, Al/Si/Ca,
Br/Pb
PM variables used:
Tracers:, I, SCU.V, Pb
Other:, EC, OC
Results: Increased percent of PMNs in BALF in CAPs-exposed rats at 24 h. CAPs affected lung tissue mRNA involved in pro-inflammation,
immune, and vascular endothelial responses. Linear regression: Increased PMN associated with CAPs mass, Br, Pb, SO4, EC, and OC.
Reference:
Nikolov et al
(2008)
Location:
Boston, MA
Subjects: dogs
Exposure:
n: 8 dogs, 24 exposure-
days in 1997-98; 4 dogs, 21
exposure-days in 2001-02
Constituents considered
for grouping: 13 elements,
BC, EC, OC
Grouping method:
Compared 3 factor-analytic
models within a SEM model
# of groups: 4
Groups/ Factors/
Sources: Oil Combustion
V/Ni; Power Plants S
Road dust Al/Si
Motror vehicles BC/OC/EC
PM variables used: mass
contribution of every
constituent
Results: Univariate respones for respiratory outcomes; Road dust and oil combustion assoicated with decreased respiratory frequency; motor
vehicles associated with increased respiratory frequency. Motor vehicles assoicated with increased PEF. Road dust associated with decreased
penh and motor vehicles associated with increased penh. Multivariate responses for respiratory outcomes; Road dust associated with decrased
respiratory rate. Motor vehicles associated with increased airway irritation.
Reference:
Saldiva et al.
(2002)
Location:
Boston, MA
Subjects: rats (Sprague
Dawley
Exposure: CAPs (3-day
ave. mass concentration
range 126.1-481 ug/m3)
n: 7-10 rats/group
(air/sham, S02/sham,
air/CAP, SO2/CAP) x 6 runs
in different seasons
Constituents considered
for grouping: 15 elements
(used Clarke 2000 to pick
out tracers)
Grouping method:
Previous study in same city
(Clarke et al.)
Groups/ Factors/
Sources: V/Ni
S
# of groups: 9 independent
variables in univariate
regression (mass and 8
elements)
Br/Pb
Na/Cf"
Cr
PM variables used:
Tracers:
Si
S04
V
Pb
Br
CI
Other:
EC, OC
Results: Increased percent and number of PMN in majority of air and SO2 rats exposed to CAPs, but significance levels not provided.
Other responses (protein, LDH, NAG) were variable and depended upon the CAPs exposure. No CAPs effect on histopathology. Linear
regression: V, Br, Pb, SO4, EC, OC, Si, CAP mass associated with increased PMN and lymphocytes for normal+CB rats. Only V not associated
with PMN in normal rats. Lymphocyte response due to CB rats, but not observed for SO4, Si, or mass in this group. Br, Pb, SO4, EC, OC, Si
assocaited with increased total protein in CB rats. CI and V associated with decreased LDH in CB rats. No BAL effects in normal rats exposed to
CAPs. V, Br, Pb, EC, OC, and CI associated with increased neutrophil density in lungs of normal rats.
Reference:
Wellenius et
al. (2003)
Location:
Boston, MA
Subjects: dogs
Exposure: CAPs (ave.
mass concentration range
161.3-957.3 ug/m3)
repeated exposure with
several weeks in between
n:6 dogs, 20 exposures
Constituents considered
for grouping: 15 elements
(+EC OC?) (taken from
Clarke et al. 2000)
Grouping method: 8
independent variables in
univariate regression
(mass, number, and 6
elements)
4 in the multivariate step
# of groups: x
Groups/ Factors/
Sources: V/Ni
S
Al/Si
Br/Pb
Na/CI
Cr
PM variables used:
Univariate:
Mass
Number
Ni, S, Si, BC
Multivariate:
Ni, S, Si, BC
Results: ST-segment elevation increased with CAPs.
Univariate regression: Si and Pb associated with peak ST-segment elevation and integrated ST-segment change.
CAPs mass or number concentration were not associated with any change.
Multivariate regression: Si associated with peak ST-segment elevation and integreated ST-segment change.
Reference:
Lippmann et
al. (2005b)
Location:
rural location
upwind from
NYC
Subjects: mice (C57 and
ApoE)
Exposure: CAPs (ave.
mass concentration 113
ug/m3)
n: C57: 3-6 mice/group
ApoE: 9-10 mice/group
Constituents considered
for grouping: 19 elements
+ OC, EC, NOs
Grouping method:
(Absolute) PCA
# of groups: 4
Groups/ Factors/
Sources: Regional SO4
(S/Si/OC). Resuspended
soil (CA/Fe/AI/Si)
RO power plants (V/Ni/Se)
Traffic and unknown
PM variables used: mass
contribution of sources
Results: ApoE null mice: Reuspended soil associated with decreased HR during exposure, but increased HR after exposure. Secondary sulfate
associated with decreased HR after exposure. Resdiual oil associated with increased RMSSD and SDNN in afternoon following exposure.
Secondary sulfate associated with decreased RMSSD and SDNN in night following exposure. Resuspended soil associated with increased
RMSSD at night following exposure. PM mass associated with decreased HR during exposure and decreased RMSSD at night following
exposure. C57 mice: Motor vehicle/other source category associated with decrease in RMSSD in afternoon following exposure
December 2008
6-253
DRAFT-DO NOT QUOTE OR CITE
-------
Reference: Subjects: NR
Maciejczyk Exposure; CAPs
poos) (90000/wel1;300 u9/ml)
Location:
rural;upwind
from NYC
n: 110 samples
Constituents considered
for grouping: 19 elements
+ OC, EC, NOs
Grouping method:
(Absolute) PCA
# of groups: 4
Groups/ Factors/
Sources: Regional SO4
Soil; Unknown
Oil combustion
PM variables used: mass
contribution of sources
Results: Correlation: V and Ni positively correlated with NF-kappaB. Oil combustion corrrelated the greatest with NF-kappaB (0.316).
Significance not provided. Only 2% of mass contribution originates from this source.
Reference:
Seagrave et
al. (2006)
Location: 4
SE US sites
for 2
Reference:
Veranth et al.
(2006)
Location: 8
sites in the
Western US
Reference:
Duvall et al.
(2008)
Location: 5
US cities
Reference:
Gurgueira et
al. (2002)
Location:
Boston, MA
Subjects rats (Fisher 344)
Exposure: 0.75,1.5 and 3
mg/rat via intratracheal
instillation
n: 5 rats/dose
Constituents considered
for grouping: NR
Grouping method: CMB:
secondary NO3; secondary
NPU; secondary SO4; coke
production; vegetative detri-
tus; natural gas combust;
road dust; wood combust;
meat cooking gasoline;
diesel other OM; other
mass
# of groups: 13
Groups/ Factors/
Sources: NR
PM variables used: mass
contribution of every
constituent, then mass
contribution of sources
Results: Potency depended upon season and site of sample collection. In general, effects were greater in the winter. PLS analysis: 2 major
constituents identified (OC, Pb, hopanes/steranes, nitrate, As for first and major metal oxides for the second), gasoline most important predictor
for both constituents, with diesel influcencing second constituent and nitrate influencing first constituent. First constituent affected cytotoxic
responses, second constituent affected inflammatory responses.
Subjects BEAS-2B cells
(35000 cells/cm2; 10, 20,
40, 80 ug/cm2)
Exposure: Loose surface
soil sweepings through
mechanical tumbler and
cascade impactor
n: 6;16 runs over 6
months.
Constituents considered
for grouping: 13 elements,
TC, 5 OC variables, 4 EC
variables, 2 ions, EU, one
ratio (Ca: Al), OP, C03
Grouping method: PLS
# of groups: NR
Groups/ Factors/
Sources: NR
PM variables used: mass
contribution (?) of every
Results: Dose-related increase in IL-6 and decreases in cell viability for all soil types. IL-8 responses more variable and dependent upon soil
type. Univariate correlations. Low correlations for all constituents tested with IL-6. Highest correlations for EC1 (R2 = 0.50) and pyrolyzed OC
(R2 = 0.46), then Ca/AI (R2 = 0.21). Carbonate carbon, EC3, and Sr correlated with IL-8 (R2 = 0.27, 0.13, and 0.25, respectively). EC and Ni
correlated with IL-8 trend over the range of 10-80 ug/cm2 (R2 = 0.39 and 0.27, respectively). Multivariate redundancy analysis OC1, OC3, OC2,
EC2, Br, EC1, Ni correlated with IL-8 release, decreased viability, and decreased IL-6 at low and high doses. Ni, EC1, and EC2 correlated with
IL-6 release at the high dose, decreased IL-6 at the low dose, decreased IL-8 release, and decreased viability. Br was negatively associated.
Subjects primary human
airway epithelial cells
(100,000 cells/ml; dose not
provided)
Exposure: NR
n: NR
Constituents considered
for grouping: NR
Grouping method: CMB,
but not on coarse and
ultrafine
# of groups: 6 or 7
Groups/ Factors/
Sources: Mobile, residual,
oil, wood, soil, secondary
SO4, secondary NO3
PM variables used: mass
contribution of sources
Results: Linear regression with individual constituents: Sulfate associated with increased IL-8 mRNA expression. Sr associated with
increased COX-2 and decreased HO-1 mRNA expressions. K associated with decreased HO-1 mRNA expression. Linear regression with
source categories: Only R2 values provided; significance levels not provided.
Subjects rats (Sprague
Dawley)
Exposure: CAPs (ave.
mass concentration 600
ug/m3); also carbon black
and ROFA
n: 13 experiments (1
rat/group at each time
point)
Constituents considered
for grouping: 20 elements
Grouping method: not
performed
# of groups: NAx
Groups/ Factors/
Sources: mass
contribution of every
constituent
PM variables used: NR
Results: Increased oxidative stress in heart and lungs following CAPs exposure (and ROFA exposure). Univariate regression: Mn, Zn, Fe, Cu,
and Ca most significant responses for lung (r2>0.40). Al, Si, Ti, Fe, and total mass most significant response for heart (r2>0.49).
Reference:
Rhoden et al.
(2004)
Location:
Boston, MA
Reference:
Kodavanti et
al. (2005)
Location:
RTP, NC
Subjects rats (Sprague
Dawley)
Exposure: CAPs (avg.
mass concentration range
150-2520 ug/m3)
acetylcysteine full factorial
n: 4-8 rats (1-2 per group -
sham, CAPs, sham/NAC,
CAP/NAC)
10 exposures
Constituents considered
for grouping: Boston, MA
Grouping method: 20
elements
# of groups: grouping not
performed
Groups/ Factors/
Sources: NR
PM variables used: mass
contribution of every
constituent
Results: Increased oxidative stress and inflammation in lungs of CAPs animals that was attenuated with NAC. Univariate regression: Al, Si, Fe,
K, Pb, and Cu most significantly correlated with lung TBARS. No significant correlations for lung carbonyls or lung PMN.
Subjects rats (SH and
WKY)
Exposure: CAPs
(144-2758 |xg/m3)
n: 6 1-day , 1 -strain runs, 7 Grouping method: Not
2-day, 2-strain runs, 4-9 performed
rats per run. # of groups: NR
Constituents considered
for grouping: NR
Groups/ Factors/
Sources: mass
contribution of every
constituent
PM variables used: NR
Results: No significant correlations between biologic responses and exposure variables (i.e., CAP mass, OC, inorganic C, sulfate, and other
major elemental constituents). Al, Cu, Zn correlated with biologic responses when constitents normalized per unit mass of CAP (ug/mg). Zn
correlated with plasma fibrinogen in SH rats (P = 0.0023).
December 2008
6-254
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
6.6.3. Summary
Recent epidemiologic, human clinical and toxicological studies have begun to evaluate the health
effects associated with ambient PM constituents and sources, as opposed to PM mass. This evaluation is
conducted using a variety of quantitative methods applied to the full set of PM constituents, rather than
selecting constituents a priori. As shown in Table 2-1, over 19 individual PM constituents have been
linked to cardiovascular and/or pulmonary responses. Similarly, a number of different PM, primarily fine
PM, source categories have been associated with health effects, including crustal and soil, traffic, second-
dary sulfate, power plants, and oil combustion. These varying results underscore the difficulty of the task.
Overall, there is no consistent trend or pattern that links particular constituents or sources with specific
health outcomes, but a number of PM2 5 constituent groupings that are commonly associated with sources
such as crustal/soil, secondary sulfate/long-range transport, traffic, oil combustion and
woodsmoke/vegetative burning were linked with health effects.
Comparisons among these studies are difficult because of differences in handling the data, in
approaches used to model source contributions, and in the level of experience in applying source appor-
tionment techniques among research groups. In an intercomparison study, several research groups used
the same data sets (which contained the composition of ambient PM2 5 and daily mortality counts) and
their choice of source apportionment models to identify PM sources (see Section 6.5.2.6). In these stu-
dies, when examining the association between various PM sources and mortality risk estimates, it was
found that the between source category variation in risk estimates for daily mortality was significantly
larger than the between group variation in reported risks. The results of this exercise indicated that the
choice of source apportionment models has a much smaller effect on variations in risk estimates com-
pared with the variations in risk caused by the different source components. Further, the most strongly
associated source types were consistent across all groups. This study indicates that source apportionment
methods can add useful insights into those source components that contribute to PM2 5 health effects.
In terms of establishing linkages between PM constituents or sources and health effects, additional
studies that increase the number of different geographic locations while examining similar health
endpoints or outcomes may help uncover a trend or pattern. In addition, the integration of results from
these types of studies would be less difficult if the methods employed for grouping PM constituents
across studies and disciplines were more consistent.
December 2008
6-255
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Chapter 7. Integrated Health Effects of
Long-Term PM Exposure
7.1. Introduction
This chapter summarizes, reviews and integrates the evidence on relationships between health
effects with long-term exposures to various size fractions and sources of PM. Cardiopulmonary health
effects of long-term exposure to PM have been examined in an extensive body of epidemiologic, human
clinical, and toxicological studies. Both epidemiologic and toxicological studies provide a basis for
examining reproductive and developmental and cancer health outcomes with regard to long-term
exposure to PM. In addition, there is a large body of epidemiologic literature evaluating the relationship
between mortality and long-term exposure to PM.
Conclusions from the 2004 PM AQCD are summarized briefly at the beginning of each section,
and the evaluation of evidence from recent studies builds upon what was available during the previous
review. For each health outcome (e.g., respiratory infections, lung function), results are summarized for
studies from the specific scientific discipline, i.e., epidemiologic, human clinical, and toxicological
studies. The sections conclude with summaries of the evidence on the various health outcomes and
integration of the findings that leads to conclusions regarding causality based upon the framework
described in Chapter 1. Determination of causality is made for the overall health effect category, such as
cardiovascular effects, with coherence and plausibility being based upon the evidence from across
disciplines and also across the suite of related health outcomes. In these summary sections, the evidence is
summarized and independent conclusions drawn for relationships with PMi0, PMi0.2.5, PM2 5, and ultrafine
particles.
7.2. Cardiovascular and Systemic Effects
Studies examining associations between long-term exposure to ambient PM (over months to years)
and CVD morbidity were not included in the 1996 or 2004 PM Air Quality Criteria Documents
(U.S. EPA, 1996, 2004), and only one study of this type (Kunzli et al., 2005) was included in the July
2006 Provisional Assessment for PM (U.S. EPA, 2004). There were no toxicological studies presented in
the 2004 PM AQCD that evaluated chronic atherosclerotic effects of PM exposure in animal models.
However, a subchronic study that evaluated atherosclerosis progression in hyperlipidemic rabbits was
December 2008
7-1
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
discussed and this study provided the foundation for the subsequent work that has been conducted in this
area (Suwa et al., 2002). No previous toxicological studies evaluated effects of subchronic or chronic PM
exposure on HR or HRV changes nor were there animal toxicological studies included in the 2004 PM
AQCD that evaluated systemic inflammatory or blood coagulation markers following subchronic or
chronic PM exposure.
Several new epidemiologic studies have examined the long-term PM-CVD association among U.S.
and European populations. The studies investigate the association of both PM2 5 and PMi0 exposures with
a variety of clinical and subclinical CVD outcomes. Epidemiologic and toxicological studies have
provided evidence of the adverse effects of long-term exposure to PM2 5 on cardiovascular effects,
including atherosclerosis and clinical and subclinical markers of cardiovascular morbidity. The evidence
of these effects from long-term exposure to PMi0 and PM10-2.5 is weaker.
7.2.1. Atherosclerosis
Atherosclerosis is a chronic inflammatory disease that contributes to several adverse outcomes,
including myocardial infarction and aortic aneurysm. It is a multifaceted disease, beginning with an early
injury or inflammation that promotes the extravasation of inflammatory cells. Under conditions of
oxidative or nitrosative stress and high lipid or cholesterol concentrations, the vessel wall undergoes a
chronic remodeling that is characterized by the presence of foam cells, migrated and differentiated
smooth muscle cells, and ultimately a fibrous cap. The advanced lesion that develops from this process
can occlude perfusion to distal tissue, causing ischemia, and erode, degrade, or even rupture, revealing
coagulant initiators (tissue factor) that cause major clotting disorders and infarction or stroke. Several
detailed reviews of atherosclerosis pathology have been published elsewhere (Ross, 1999; Stacker and
Keaney, 2004).
7.2.1.1. Epidemiologic Studies
Measures of Atherosclerosis
There are four preclinical markers that have been used in the epidemiologic studies of
atherosclerosis. These measures are described briefly below.
CAC is a measure of atherosclerosis assessed by non-contrast, cardiac-gated electron beam
computed tomography (EBCT) or multidetector computed tomography (MDCT) of the coronary arteries
in the heart (Greenland and Kizilbash, 2005; Hoffmann et al., 2005; Mollet et al., 2005). The prevalence
of CAC is strongly related to age. Few people have detectable CAC in their second decade of life;
December 2008
7-2
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
however, the prevalence of CAC rises to approximately 100% by age 80 (Ardehali et al., 2007). Previous
studies suggest that while the absence of CAC does not rule out atherosclerosis, it does imply a very low
likelihood of significant arterial obstruction (Achenbach and Daniel, 2001; Arad et al., 1996; Shaw et al.,
2003; Shemesh et al., 1996). Conversely, the presence of CAC confirms the existence of atherosclerotic
plaque and the amount of calcification varies directly with the likelihood of obstructive disease (Ardehali
et al., 2007). CAC is often quantified using the Agatston method (1990). Agatston scores are frequently
used to classify individuals into one of five groups (zero; mild; moderate; severe; extensive) or according
to age- and sex-specific percentiles of the CAC distribution (Erbel et al., 2007).
CIMT is a measure of atherosclerosis assessed by high-resolution, B-mode ultrasonography of the
carotid arteries in the neck, the walls of which have inner (intimal), middle (medial) and outer
(adventitial) layers (Craven et al., 1990; O'Leary et al., 1999; Wendelhag et al., 1993). CIMT estimates
the distance in mm or |_im between the innermost (blood-intima) and outermost (media-adventitia)
interfaces, often by averaging over three arterial segments in the common carotid, carotid bulb, and
internal carotid artery (Amato et al., 2007). CIMT has been associated with atherosclerosis risk factors
(Heiss et al., 1991; O'Leary et al., 1992; Salonen and Salonen, 1991), prevalent coronary heart disease
(Chambless et al., 1997; Geroulakos et al., 1994), and incident coronary and cerebral events (O'Leary et
al., 1999; van der Meer et al., 2004). Several studies have indicated that CIMT measurements are accurate
(Girerd et al., 1994; Pignoli et al., 1986; Wendelhag et al., 1991) and reproducible (Montauban van
Swijndregt et al., 1999; Smilde et al., 1997; Willekes et al., 1999), perhaps most so for the common
carotid artery (Montauban van Swijndregt et al., 1999).
ABI—also known as the ankle-arm or resting (blood) pressure index—is a measure of lower
extremity arterial occlusive disease commonly caused by advanced atherosclerosis (Weitz et al., 1996). It
is assessed by continuous wave Doppler and manual or automated oscillometric sphygmomanometry, the
latter having been shown to have higher repeatability and validity (Weitz et al., 1996). ABI is defined as
the unitless ratio of ankle to brachial systolic blood pressures measured in mm Hg. As ankle pressure is
normally equal to or slightly higher than arm pressure (resulting in an ABI > 1.0), epidemiologic studies
typically define the normal ABI range as 0.90 to 1.50 (Resnick et al., 2004). Low ABI has been associated
with all-cause and CVD mortality (Newman et al., 1993; Vogt et al., 1993), as well as myocardial
infarction and stroke (Karthikeyan and Lip, 2007).
AAC is a measure of atherosclerosis assessed by non-contrast, EBCT or MDCT of the abdominal
aorta. It is scored much like CAC (Agatston et al., 1990), but the age-specific prevalence and extent of
AAC is greater, particularly among women and at ages >50 years. Although AAC has not been studied as
extensively as CAC, it is associated with carotid and coronary atherosclerosis as well as cardiovascular
December 2008
7-3
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
morbidity and mortality (Allison et al., 2004; 2006; Hollander et al., 2003; Khoury et al., 1997; Oei et al.,
2002; Walsh et al., 2002; Wilson et al., 2001; Witteman et al., 1986).
Study Findings
Two studies examined the long-term PM-CAC association (Diez Roux et al., 2008; Hoffmann et
al., 2007). Diez Roux et al. (2008) studied 5,172 residents of Baltimore, MD; Chicago, IL; Forsyth Co,
NC; Los Angeles, CA; New York, NY; and St. Paul, MN (age range 45-84 yr; 53% female) at the
baseline exam of the MESA (2000-2002). In this cross-sectional ancillary study, the authors used spatio-
temporal modeling of pollutant concentrations, National Climatic Data Center climate, and U.S. Census
demographic data to impute 20-year average exposures to PM2 5 and PMi0. They found that 10 |ig/m3
increases in PMi0 and PM2 5 were associated with 1% (95% CI: -2 to 4) and 0.5% (95% CI: -2 to 3)
increases in the relative prevalence of CAC, respectively. Among the subset of 2,586 participants with
EBCT-identified calcification, corresponding increases in CAC were 0.5% (95% CI: -7 to 9) and 0.5%
(95% CI: -5 to 7) or approximately 1 (95% CI: -10 to 13) and 1 (95% CI: -13 to 16) Agatston units,
respectively. There was little evidence of effect modification by demographic, socioeconomic or clinical
characteristics.
Hoffman et al. (2007) studied 4,494 residents of Essen, Miilheim and Bochum, Germany (age
range: 45-74 yr; 51% female) at the baseline exam of the Heinz Nixdorf Recall Study (2000-2003). In
this cross-sectional study the authors used dispersion modeling of emissions, climate and topography data
to estimate one-year average exposure to PM2 5. They found that a 10 |ig/m3 increase in PM2 5 was
associated with a 43% (95% CI: 15-115) or 102 (95% CI: 77-273) Agatston increase in CAC. Differences
in strength of association between subgroups defined by demographic and clinical characteristics were
small.
Two studies examined the chronic PM-CIMT association (Diez Roux et al., 2008; Kunzli et al.,
2005). Diez Roux et al. (2008) used cross-sectional data from the MESA cohort (described previously)
and Kunzli et al. (2005) used cross-sectional data from the the Vitamin E Atherosclerosis Progression
Study (VEAPS) and B-Vitamin Atherosclerosis Intervention Trial (BVAIT).
Diez Roux et al. (2008) found that 10 |ig/m3 increases in 20-year average PMi0 and PM2 5
concentrations were associated with 1% (95% CI: 0-1.4) and 0.5% (95% CI: 0-1) or approximately
8 (95% CI: 0-12) and 7 (95% CI: 0-14) |_im increases in CIMT, respectively. Evidence of age-, gender-,
lipid-and smoking-related susceptibility was lacking in this context.
Kunzli et al. (2005) studied 798 residents of the greater Los Angeles, CA area (age >40 yr; 44%
female) at the baseline exams of two randomized, placebo-controlled clinical trials (VEAPS and BVAIT,
1998-2003). In this cross-sectional ancillary study of these extant cohorts, the authors used universal
December 2008
7-4
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
kriging of PM25 data from 23 state and local monitors operating in 2000 to estimate 1-year average
exposure to PM2 5 at each participant's geocoded U.S. Postal Service ZIP code. They found that a
10 |ig/m3 increase in exposure was associated with a 4.2% (95% CI: -0.2 to 8.9) or approximately 32
(95% CI: -2 to 68) (jm increase in CIMT. In contrast to findings from the relatively large, ethnically
diverse, yet geographically overlapping MESA ancillary study described above, PM-related increases in
CIMT were two- to three-fold larger among older and female participants taking anti-hyperlipidemics in
this study. They were also higher in never than in current or former smokers.
In addition to examining the PM-CIMT association, Diez Roux et al. (2008) examined the chronic
PM-ABI association (Diez Roux et al., 2008). The authors found that 10 |ig/m3 increases in 20-year
average PMi0 and PM25 concentratins were associated with 0 (95% CI: -0.003 to 0.004) and -0.001 (95%
CI: -0.005 to 0.005) mean differences in ABI, respectively. These largely null findings exhibited little
heterogeneity among participant subgroups and were similarly null when ABI was modeled as a
dichotomous outcome using a cutpoint of 0.9 units.
One study examined the chronic PM25-AAC association in a residentially stable subset of 1,147
participants (mean age = 66 yr; 50% female) randomly selected from all MESA centers (except
Baltimore, MD) for enrollment in its Aortic Calcium Ancillary Study (Allen et al., in press). The authors
used kriging and inverse residence-to-monitor distance-weighted averaging of EPAAQS data to estimate
two-year mean exposures to PM2 5. In cross-sectional analyses, the authors found that 10 |ig/m3 increases
in PM2 5 exposures were associated with 6% (95% CI: -4 to 16) excess risk of AAC and 8% (95%
CI: -30 to 46) increases in AAC, i.e. approximately 50 (95% CI: -251 to 385) Agatston units. These
associations were stronger among users than non-users of anti-hyperlipidemics.
The ambient PM concentrations from these studies are characterized in Table 7-1.
7.2.1.2. Toxicological Studies
In the only study of this kind described in the 2004 PM AQCD, Suwa et al. (2002) conducted an
experiment to evaluate atherosclerosis progression in rabbits exposed to PMi0. The rabbits exposed to
PM10 (5 mg/kg, 2 times/wk x 4 wk) demonstrated more advanced atherosclerotic lesions based on
phenotype and volume fraction in the left main and right coronary arteries. More extensive atherosclerosis
was also observed in the aorta of PMio-exposed animals, with increased extracellular lipid pools and total
amount of lipids in the lesions. Although this study was conducted at a relatively high dose, it provided
the first experimental evidence that PM exposure may result in progression of atherosclerosis.
December 2008
7-5
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
CAPs
Sun et al. (2005) conducted a 6-month exposure of ApoE"7" mice to Tuxedo, NY CAPs from July
2004 to January 2005 and demonstrated an enhancement of atherosclerosis. Mice were exposed to PM2 5
concentrated to roughly 85 |ig/m3 for 6 h/d><5 d/wk; the average exposure concentration when normalized
over the entire 6 month period (24 h/d><7d/wk) was 15.2 (.ig/ni1. PM2 5 exposure had no effect on the net
concentrations of cholesterol or triglycerides. Plaque area in the aortic arch and abdominal aorta was
significantly increased in ApoE"" mice exposed to CAPs on a high-fat diet compared to filtered air mice
fed high-fat chow (41.5% and 26.2%, respectively), and lipid content in the thoracic aorta reflected the
plaque area response with the PM-exposed high fat-chow group having greater lipid staining vs. the
control group on a high fat diet (30.0 and 20.0, respectively). Macrophage infiltration in the abdominal
aorta (primarily in the intimal and medial areas) was also observed in the groups exposed to CAPs.
Significant enhancement of 3-nitrotyrosine and inducible NOS were observed in the abdominal aorta in
the PM2 5-exposed mice for the normal and high-fat chow groups. Furthermore, aortas from the PM2 5-
exposed animals exhibited increased vasoconstrictor responsiveness to serotonin and reduced dilatation to
acetylcholine.
A recent study (Sun et al., 2008) that was part of the research described above (Sun et al., 2005)
investigated tissue factor (TF) expression in aortas, which is a major regulator of hemostasis and
thrombosis following vascular injury or plaque erosion. In PM2 5-exposed ApoE"" mice on a high-fat diet,
TF was significantly elevated in the plaques of aortic sections compared to air-exposed mice on the high-
fat diet. There was no difference in TF immunohistochemistry staining in the CAPs-exposed or control
groups fed normal chow. TF expression was generally detected in (1) the extracellular matrix surrounding
macrophages and foam cell-rich areas and (2) around smooth muscle cells. Vascular macrophage
infiltration was increased with PM2 5 exposure independent of diet and was found predominantly in the
intimal surface and within the plaque.
Several other projects have examined the subchronic effects of PM and other air pollutants on
ApoE"'" mice. Chen and Nadziejko (2005) investigated histopathological changes in the aortas of both
ApoE" " mice and the double-knockout ApoE7 /LDLR7 mice following a 4-5 month (March, April or May
through September 2003) exposure to Tuxedo, NY CAPs. The average PM2 5 exposure concentration
ranged from 110 to 131 (.ig/ni1. This study reported increased mortality in the double-knockout mice
exposed to air (n = 9) and CAPs (n = 11) and it appeared that the CAPs-exposed animals died earlier than
the air controls. At the end of exposure, the number of double-knockout mice with coronary artery disease
was greater in the CAPs-exposed group, with 7 of 10 having some degree of lipid deposition (compared
to 3 of 13 for the air group). The enhancement of aortic lesion area was consistent with the findings of
Sun et al. (2005) and both lesion area and lesion cellularity were enhanced by CAPs exposure in the male
December 2008
7-6
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
double-knockout mice, although there was no change in lipid content and there were no differences in
grossly discernible plaque with mice on the high-fat diet. The percentage of aortic intimal surface covered
by atherosclerotic lesions in ApoE" " mice was increased by 57% in the CAPs-exposed group compared to
the air-exposed group.
One new study of fine or ultrafine PM derived from traffic was conducted. Araujo et al. (2008)
compared the relative impact of ultrafine (0.01-0.18 |im) and fine (0.01-2.5 |im) PM inhalation on aortic
lesion development in ApoE_/" mice following a 40-day exposure (5 h/d><3 d/wk for 75 total h). Animals
were on a normal chow diet and exposed to CAPs from November through December 2005 in a mobile
inhalation laboratory that was parked 300 m from the 110 Freeway in downtown Los Angeles. Particles
were concentrated to -440 |ig/m3 for the fine exposures and to ~110 |ig/m3 for the ultrafine exposures,
representing a roughly 15-fold concentration from ambient levels; the number concentration of PM in the
fine and ultrafine chambers were roughly equivalent (4.56/105 and 5.59/105 particles/cm3, respectively).
The authors noted significant increases in plaque size (estimated by lesions at the aortic root) in the
ultrafine PM-exposed mice compared to both control and fine PM-exposed, with no difference observed
between control and fine PM-exposed mice. The lesions were largely comprised of macrophage
infiltration with intracellular lipid accumulation. Increased total cholesterol measured at the end of the
exposure protocol was observed only in the fine PM group. High density lipoprotein isolated from the
ultrafine PM-exposed mice demonstrated decreased HDL anti-inflammatory protective capacity against
LDL-induced monocyte chemotactic activity in an in vitro assay. The livers from the ultrafine
PM-exposed mice demonstrated significant increases in lipid peroxidation and several stress-related gene
products (catalase, glutathione S-transferase Ya, NADPH-quinone oxidoreductasel, superoxide
dismutase 2). Thus, ultrafine PM in these exposures had a substantially greater impact on the systemic
response than did fine PM, despite 85% of the total particle number concentration for PM2 5 being
comprised of ultrafine PM. The authors suggested that the proportional difference in organic carbon
composition of the two atmospheres or increased particle surface area may be responsible for the
observed differences in biological outcome.
PM10
A study employing young BALB/c mice examined the effects of a 4 month exposure (24 h/d x
7 d/wk) to ambient air on arterial histopathology (Lemos et al., 2006). Outdoor exposure chambers were
located in downtown Sao Paulo, Brazil next to streets of high traffic density and the gases were not
filtered. In the control chamber, PMi0 and N02 were filtered with 50% and 75% efficiency, respectively.
The average pollutant concentrations were 2.06 ppm for CO (8-h mean), 104.75 (ig/m3 for N02 (24-h
mean), 11.07 (.ig/nr3 for S02 (24-h mean), and 35.52 (.ig/nr3 for PMi0 (24-h mean) at a monitoring site
December 2008
7-7
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
within 100 m of the inhalation chambers. The pulmonary and coronary arteries demonstrated significant
decreases in lumen/wall (LAV) ratio for animals exposed to the entire ambient mixture compared to
controls, indicating thicker walls in these vessels. There was no difference reported for the LAV ratio in
renal arteries. Morphologic examination suggested that the increases in LAV ratio were due to muscular
hypertrophy rather than fibrosis. The results of this study indicate vascular remodeling of the pulmonary
and coronary arteries, as opposed to changes in tone.
To examine the role of systemic inflammation and recruitment of monocytes into plaque tissue as a
possible pathway for accelerated atherosclerosis, Yatera et al. (2008) exposed female Watanabe heritable
hyperlipidemic rabbits (42 wks old) to Ottawa PMi0 (EHC-93) via intratracheal instillation (5 mg/rabbit;
approximately 1.56 mg/kg) twice a week for 4 weeks. Transfusion of whole blood harvested to from
exposed and non-exposed animals to donor rabbits supplied labeled monocytes for assessment of
monocyte recruitment from the blood to the aortic wall. The fraction of aortic surface and volume of
aortic wall taken up by atherosclerotic plaque was increased and the number of labeled monocytes in the
atherosclerotic plaques was elevated in rabbits exposed to PMi0. In addition, labeled monocytes were
attached onto the endothelium overlying atherosclerotic plaques and the number that migrated into the
smooth muscle underneath plaques in aortic vessel walls was greater with PMi0 exposure compared to
control. These responses were not observed in normal vessel walls. ICAM-1 and VCAM-1 expression
was elevated in atherosclerotic lesions, likely indicating enhanced monocyte adhesion to endothelium and
migration into plaques. Monocytes in plaque tissue stained with immunogold demonstrated foam cell
characteristics, which were more numerous in the rabbits exposed to PM10.
Gasoline Exhaust
Lund and colleagues (2007) used whole emissions from gasoline exhaust to investigate changes in
the transcriptional regulation of several gene products with known roles in both the chronic promotion
and acute degradation/destabilization of atheromatous plaques. These 50-day exposures (6 h/d><7 d/wk)
employed ApoE" " mice on high-fat chow and the concentrations of the high exposure group were
61 |ig/m3 for PM, 19 ppm for NOx, 80 ppm for CO, and 12.0 ppm for total hydrocarbons. The average
particle number median diameter was approximately 15 nm (McDonald et al., 2007). Dilutions of
gasoline engine emissions induced a concentration-dependent increase in transcription of matrix
metalloproteinase (MMP) isoform 9, ET-1, and HO-1 in aortas; MMP-3 and -9 mRNA levels were only
increased in animals in the highest exposure group. Strong increases in oxidative stress markers
(nitrotyrosine and TBARS) in the aortas were also observed. However, using a high-efficiency particle
trap, they established that most of the effects were caused by the gaseous portion of the emissions and not
the particles. This study did not directly address lesion area and is not directly comparable to the studies
December 2008
7-8
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
by Sun et al. (2005), Chen and Nadziejko (2005), and Araujo et al. (2008), although it suggests that the
gases are important contributors to overall toxicity.
7.2.2. Thromboembolism
The relationship between PM exposure and health outcomes indicative of thrombosis or embolism
formation was evaluated in one epidemiologic study. Prothrombin and partial thomboplastin times (PT;
PTT) are laboratory measures of hemostasis. PT and PTT measure the extrinsic and intrinsic blood
coagulation pathways, the former activated in response to blood vessel injury, the latter, key to subsequent
amplification of the coagulation cascade and propagation of thrombus (Mackman et al., 2007).
7.2.2.1. Epidemiologic Studies
Baccarelli et al. (2007b) studied 2,081 residents of the Lombardy region of Italy (age range 18-84
yr; 56% female). In this case-control study of 871 patients with ultrasonographically or venographically
diagnosed lower extremity deep vein thrombosis (DVT) and 1,210 of their healthy friends or relatives
(1995-2005), the authors used arithmetic averaging of PMi0 data available at 53 monitors in nine
geographic areas to estimate one-year average residence-specific exposures. They found that a 10 |ig/m3
increase in PM10 was associated with -0.09 (95% CI: -0.16 to 0) and -0.18 (95% CI: -0.35 to 0) decreases
in standardized correlation coefficients for PT as well as 0.02 (95% CI: -0.05 to 0.06) and -0.11 (95%
CI: -0.29 to 0.12) decreases in standardized correlation coefficients for activated partial thromboplastin
time (aPTT) among cases and controls, respectively. Shortened PT and aPTT reflect hypercoagulability.
However, patients with DVT who were taking heparin or coumarin anticoagulants were not asked to stop
taking them before measurement of PT and aPTT. Moreover, PT was neither adjusted for differences in
reagents used to determine it nor conventionally reported as the International Normalized Ratio (INR).
The ambient PM concentrations from this study are characterized in Table 7-1.
7.2.3. Systemic Inflammation and Blood Coagulation
7.2.3.1. Toxicological Studies
In addition to the PM2 5 study above that showed increased TF expression (an important initiator of
thrombosis) in aortas of ApoE" "following subscronic CAPs exposure (Sun et al., 2008), three recent
studies conducted by the same group examined hematology and clotting parameters in rats and mice
exposed to DE, gasoline exhaust, or HWS for 1 week or 6 months (Reed et al., 2004; 2006; 2008). In all
December 2008
7-9
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
studies, male and female F344 rats were exposed to the mixtures by whole-body inhalation for 6 h/day,
7 day/wk. The target PM concentrations in the diesel and HWS studies were 30, 100, 300, and
1000 |ig/m3 (Reed et al., 2004; 2006); the dilutions for the gasoline exhaust were 1:10, 1:15 or 1:90 and
filtered PM at the 1:10 dilution (Reed et al., 2008). PM mass in the latter study ranged from 6.6 to
59.1 |ig/m3, with the corresponding number concentration between 2.6/104 and 5.Ox 105 particles/cm3.
Respiratory effects for these studies are presented in Section 7.3.4.
Male and female rats exposed to DE at the highest concentration (NO concentration 45.3 ppm; N02
concentration 4.0 ppm; CO concentration 29.8 ppm; S02 concentration 365 ppb) for 6 months
demonstrated decreased serum Factor VII, but no change in plasma fibrinogen or TAT (Reed et al., 2004).
Together, these findings likely do not support an exposure-related stimulation of blood coagulation. White
blood cells were decreased only in female rats in the highest exposure group.
In male rats exposed to HWS, the mid-low group (PM concentration 113 |ig/m3: NO concentration
0 ppm; N02 concentration 0 ppm; CO concentration 1832.3 ppm; S02 concentration 0 ppb) had the
greatest responses in hematology parameters, including increased hematocrit, hemoglobin, lymphocytes,
and decreased segmented neutrophils (Reed et al., 2006). Platelets were elevated in male and female rats
after 1 week of exposure, but this response returned to control values following the 6 month exposure. No
changes were observed for any coagulation markers at 6 months.
Similar to the responses observed with HWS, male and female rats in the mid and high gasoline
exhaust exposure groups (NO concentrations 11.9 and 18.4 ppm; N02 concentrations 0.5 and 0.9 ppm;
CO concentration 73.2 and 107.3 ppm; S02 concentration 0.38 and 0.62 ppm, respectively) demonstrated
elevated hematocrit and hemoglobin; red blood cell count was also elevated in these groups (Reed et al.,
2008). The only response that appeared somewhat dependent on the presence of particles was increased
RBC in female rats at 6 months, although the authors attributed the observed increases to the high
concentration of CO.
Collectively, these studies do not indicate robust systemic inflammation or coagulation responses in
F344 rats following 6-month exposures to diesel, HWS, or gasoline exhaust. The limited effects that were
observed could possibly be due to the varying gas concentrations in the exposure mixtures.
7.2.4. Renal and Vascular Function
Two recent epidemiologic studies have tested associations between PM exposure and indicators of
renal (urinary albumin to creatinine ratio [UACR]) or vascular (blood pressure) function. UACR is a
measure of urinary albumin excretion (National Kidney Foundation, 2008). When calculated as the ratio
of albumin to creatinine concentrations in untimed ("spot") urine samples, UACR approximates 24-h
December 2008
7-10
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
urinary albumin excretion and can be used to identify albuminuria, a marker of generalized vascular
endothelial damage (Xu et al., 2008). Values > 30 mg/g (3.5 mg/mmol) and > 300 mg/g (34 mg/mmol)
usually define micro- and macroalbuminuria, both of which are associated with increases in CVD
incidence and mortality (Bigazzi et al., 1998; Deckert et al., 1996; Dinneen and Gerstein, 1997; Gerstein
et al., 2001; Mogensen, 1984). Several researchers have called the dichotomization of albuminuria into
question, observing that there is no threshold below which risk of cardiovascular and end-stage kidney
disease disappears (Forman and Brenner, 2006; Knight and Curhan, 2003; Ruggenenti and Remuzzi,
2006).
Systolic, diastolic, pulse, and mean arterial blood pressures (SBP; DBP; PP; MAP) in mm Hg have
also been used as measures of cardiovascular disease. Franklin et al. (1997) suggested that SBP and PP
were the only two measures predictive of carotid stenosis in a multivariable analysis considering all four
measures (Franklin et al., 1997), whereas Khattar et al. (2001) suggested that their prognostic significance
in hypertensive populations may differ by age, with SBP and PP being most predictive among those > 60
and DBP among those <60 years old (Khattar et al., 2001).
7.2.4.1. Epidemiologic Studies
One epidemiologic study examined the long-term PM-UACR association (O'Neill et al., 2007).
Although this study also was based on MESA ancillary study data described previously (Diez Roux et al.,
2008), its cross-sectional and longitudinal analyses focused on a subset of 3,901 MESA participants
(mean age = 63 yr; 52% female) with complete covariate, outcome and exposure data at their first through
third exams (2000-2004). In cross-sectional analyses, the authors found that after adjustment for
demographic and clinical characteristics, 10 |ig/m3 increases in 20-year imputed exposures to PM2 5 and
PM10 were associated with (1) 0.002 (95% CI: -0.048 to 0.052) and -0.002 (95% CI: -0.038 to 0.035)
mean differences in baseline log AUCR; and (2) -2% (95% CI: -16 to 14) and -2% (95% CI: -13 to 10)
decreases in the baseline prevalence of microalbuminuria (defined in this setting as > 25 mg/g). These
largely null cross-sectional findings mirrored those based on the study's shorter-term (30- and 60-day)
PM2 5 and PMi0 exposures. Moreover, longitudinal analyses revealed only a weak association between
three-year change in log UACR and 20-year PMi0 exposure. Evidence of effect modification by
demographic and geographic characteristics was negligible in cross-sectional and longitudinal analyses
alike.
In another study, Auchincloss et al. (2008) focused on automated, oscillometric,
sphygmomanometric measures of blood pressures in mm Hg (SBP; DBP; PP; MAP). Like O'Neill et al.
(2007), Diez Roux et al. (2008) and Allen et al (in press). Auchincloss et al. (2008) based their
examination on the previously described MESA population. The authors studied 5,112 participants (age
December 2008
7-11
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
range = 45-84 yr; 52% female) who were free of clinically manifested CVD at their baseline exam in one
of six, primarily urban U.S. locations (2000-2002). In this cross-sectional study, they used arithmetic
averaging of EPA AQS PM2 5 data available at the monitor nearest to each participant's geocoded U.S.
Postal Service ZIP code centroid to estimate 30- and 60-day average exposures to PM25. They found that
a 10 (ig/m3 increase in 60-day average PM2.5 exposure was associated with 1.20 (95% CI: -1.0 to 3.4) to -
0.1 (95% CI: -1.2 to 1.0), 1.3 (95% CI: -0.4 to 2.9), and 0.4 (95% CI: -1.01 to 1.7) mm Hg increases in
SBP, DBP, PP and MAP. Associations were slightly stronger for 30-day average PM2 5 exposure,
particularly for SBP and PP among participants with hypertension or taking anti-hypertensives, during
warmer weather, in the presence of high N02, residing < 300 m from a highway, or surrounded by higher
road density.
Finally, Calderon-Garciduenas et al. (2007) studied ET-1 and pulmonary artery pressure in two
cohorts of healthy children aged 6-13 years old in Mexico. The exposed cohort consisted of children from
two areas in Mexico City with different pollution profiles. The control group was drawn from a less
polluted area in Mexico (Polotitlan). Children enrolled in the study were lifelong residents in their
community, and, lived and attended school within 5 miles of one of the air monitoring stations used to
estimate ambient exposures. Cardiovascular function was assessed using Doppler echocardiography and
fasting blood was collected and analyzed for complete blood count and ET-1. The authors reported that
long-term exposure to air pollution (e.g. residence and school attendance in Mexico City compared to the
control city) was associated with elevated mean pulmonary arterial pressure (17.3 ± 0.5 vs. 14.6 ±
-0.4 mmHg, p<0.01) and increased ET-1 (2.24 ± -0.12 vs. 1.23 ± -0.06,/><0.001). ET-1 levels were
significantly higher in children from the Northeast section of Mexico City compared to children from the
Southwest section of the city. In order to distinguish the effect of PM2 5 from PMi0 and ozone
investigators examined the differences in pollution profiles of these two areas of Mexico City. PM2 5 level
in the Northeast was significantly higher than in the Southwest. This was not true for PMi0 and ozone
level, which was higher in the Southwest. Further, cumulative PM2 5 level 7 days prior to the blood draw
was correlated with increased levels of circulating ET-1 (r = 0.28, p = 0.03). PM concentrations from
these epidemiologic studies are characterized in Table 7-1.
December 2008
7-12
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
Table 7-1. Characterization of ambient PM concentrations from studies of subclinical measures
of cardiovascular diseases.
Reference
Location
Mean Concentration (jjg/m3)
Upper Percentile
Concentrations (|jg/m3)
PMw
Diez-Roux (2008)
MESA: 6 Cities U.S.
20 year imputed mean: 34
NR
O'Neill et al. (2007) MESA: 6 Cities U.S. Long-Term Exposure: NR
1982-2002:34.7
1982-1987: 40.5
1988-1992:38
1993-1997:30.6
1998-2002:29.7
Previous Month: 27.5
Baccarelli et al. (2008) Lombardy Italy
Rosenlund et al. (2006) Stockholm, Sweden
Annual avg: 41 NR
30-y avg PM10 (traffic) 5th-95th%:
Cases: 2.6 0.5-6
Controls: 2.4 0.6-5.9
PMis
Hoffman et al. (2007) HNRS, 3 Cities Germany Annual avg: 22.8 NR
Allen et al. (in press) MESA: 5 Cities Annual avg: 15.8 Min-Max: 10.6-24.7
Kunzli et al. (2005) VEAPS BVAIT Annual avg: 20.3 Min-Max: 5.2-26.9
Auchincloss et al. (2008)' MESA: 6 Cities Prior 30 d: 16.8 NR
Prior 60 d: 16.7
O'Neill et al. (2007) MESA: 6 Cities U.S. Previous Month: 16.5 NR
Diez-Roux et al. (2008) MESA: 6 Cities U.S. 20-y imputed mean: 21.7 NR
MESA: Multi-Ethnic Study of Atherosclerosis
HNRS: Heinz Nixdorf Recall Study
VEAPS: Vitamin E Atherosclerosis Progression Study
BVAIT: B-Vitamin Atherosclerosis Intervention Trial
WHI: Womens Health Initiative
7.2.4.2. Toxicological Studies
In a CAPs study of shorter duration (10 weeks; 6 h/d><5 d/wk) in Tuxedo, NY (mean chamber
concentration of PM2 5 of 79.1 |ig/m3). there was no difference in mean arterial pressure (MAP) in
Sprague Dawley rats between groups (Sun et al., 2008). When angiotensin II (Ang II) was infused during
the last week of exposure to induce systemic hyptertension, the MAP slope was consistently greater in the
CAPs-exposed rats compared to the filtered air group. Furthermore, thoracic aortic rings were more
responsive to phenylephrine-induced constriction and less responsive to ACh-induced relaxation in the
PM+Ang II vessels. In contrast to the latter findings, the relaxation response was exaggerated in the
PM+Ang II aortic segments with a Rho-kinase (ROCK) inhibitor. Superoxide production in aortic rings
increased in the PM+Ang II group compared to the filtered air group and the addition of NAD(P)H
oxidase inhibitor (apocymin) or a NOS inhibitor (L-NAME) attenuated the superoxide generation. The
levels of tetrahydrobiopterin (BH4) were decreased in mesenteric vasculature and the heart by 46% and
December 2008
7-13
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
41% in the PM+Ang II group compared to controls, respectively; furthermore, levels of BH4 in the liver
were similarly reduced, which is consistent with a systemic effect of CAPs. Together, these findings
indicate that CAPs potentiate Ang II-induced hypertension and alter vascular reactivity, perhaps through
activated NADPH oxidase and eNOS uncoupling that result in oxidative stress generation and triggering
of the Rho/ROCK signaling pathway.
7.2.5. Autonomic Function
Toxicological Studies
Hwang et al. (2005) and Chen and Hwang (2005) used radiotelemetry to examine the chronic
changes in HR, HRV, physical activity (PA), and temperature (Tco) resulting from the same CAPs
exposures described above (Chen and Hwang, 2005). The overall average CAPs exposure concentration
was 133 (.ig/ni1 and results indicate differing responses to CAPs between ApoE"'" mice and their genetic
background strain, C57BL/6J mice (Hwang et al., 2005). Using the time period of 1:30 to 4:30 a.m.,
C57BL/6J mice showed a 0.4°C Tco increase over the entire exposure period, with HR only demonstrating
elevations over the last month of exposure. In contrast, ApoE"" mice had chronic decreases of 1.0°C for
Tco and a 33.8 beat/min for HR. Changes in HRV (SDNN and rMSSD) were somewhat more complicated,
with biphasic responses in ApoE" " mice over the 5 month period (initial increase over first 6 wk, decrease
over next 12 wk, and slight upward turn for remainder of the study) (Chen and Hwang, 2005). Increasing
linear trends were observed in C57BL/6J mice for SDNN and rMSSD. The average CAPs concentration
for the HRV study was 110 (ig/m3. However, only 3 C57BL/6J mice in the exposure group were included
in the analysis compared to 10 ApoE" " animals, thus making it difficult to interpret the C57BL/6J mice
responses (Chen and Hwang, 2005; Hwang et al., 2005).
7.2.6. Clinical Outcomes in Epidemiologic Studies
Several epidemiologic studies of U.S. and European populations have examined associations
between long-term PM exposures and clinical CVD events (Baccarelli et al., 2008; Hoffmann et al., 2006;
Maheswaran et al., 2005a, b; Miller et al., 2007b; Rosenlund et al., 2006; Solomon et al., 2003; Zanobetti
and Schwartz, 2007). Results from these studies are summarized in Figure 7-1. The ambient PM
concentrations from these studies are characterized in Table 7-1.
December 2008
7-14
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Coronary Heart Disease (CHD)
Several studies examined the chronic PM (or BS)-CHD association (Hoffmann et al., 2006;
Maheswaran et al., 2005a; Miller et al., 2007b; Rosenlund et al., 2006; Solomon et al., 2003; Zanobetti
and Schwartz, 2007). In these studies, CHD was variably defined and included incident (versus prevalent)
disease and validated (versus hospital-coded or self-reported) history of angina pectoris, myocardial
infarction (MI), coronary artery revascularization (bypass graft; angioplasty; stent; atherectomy), and
congestive heart failure (CHF). Results pertaining to death from CHD are described in Section 7.6.
Puett et al. (2008) studied incident, validated CHD, CHD death, and non-fatal MI among 66,250
female residents (mean age = 62 yr) of metropolitan statistical areas in thirteen northeastern U.S. states
who were enrolled in the Nurses' Health Study (NHS, 1992-2002). In this prospective cohort study of
women without a history of non-fatal MI at baseline (maximum duration of follow-up = 4 yr), the authors
used two-stage, spatially smoothed, land use regression to estimate residence-specific, one-year moving
average PMi0 exposures from U.S. EPA AQS and emissions, IMPROVE, and Harvard University monitor
data. They found a 10 |i/m3 increase in PMi0 exposure was associated with a 30% (95% CI: 0 -71)
increase in CHD death. Associations with CHD death were higher in the obese and in the never smokers.
The association of PMi0 with incident CHD was non-significantly elevated and the association of PMi0
with MI was close to the null value.
Miller et al. (2007b) studied incident, validated MI, revascularization, and CHD death, both
separately and collectively, among 58,610 post-menopausal female residents of 36 U.S. metropolitan
areas (age range = 50-79 yr) enrolled in the Women's Health Initiative Observational Study (WHI OS,
1994-1998). In this prospective cohort study of participants free of CVD at baseline (median duration of
follow-up = 6 yr), the authors used arithmetic averaging of year 2000 EPA AQS PM2 5 data available at
the monitor nearest to each participant's geocoded U.S. Postal Service five-digit ZIP code centroid to
estimate one-year average exposures. They found that a 10 (ig/m3 increase in PM2.5 exposure was
associated with 6% (95% CI: -15 to 34), 20% (95% CI: 0 to 43) and 21% (95% CI: 4 to 42) increases in
the overall hazard of MI, revascularization, and their combination with CHD death, respectively. Hazards
were higher within than between cities and in the obese.
December 2008
7-15
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Location
Averaging Time
Puett et a 1.2008
Zanobetti
and Schwartz 2007
NE US
Metropolitan Areas
21 Cities US
1-| avg
l-y avg
Rosenlund etai. 2006 Stockholm. Sweden 30-y avg
Maheswaren et al. 2005 Sheffield, UK 5-y avg
Maheswaren et at. 2005 Sheffield, UK 5-y avg
Baccarellietal, 2008 Lombarely, Itaiy 1 -y av§
Miller et al, 2007 36 Cities US
1-y avg
Hoffman etal, 2006 2 Cities Germany 1-y avg
Pi 10
PM 2 5
- 1st CHD Event
- Recurrent it
Post-il CHF
Non-fatal Ml
CHD HospitaRzation
Stroke Hospitabatton
Deep Vein Thrombosis
Incident Ml
Revascufartzaftoo
Stroke
Cerebrovascular Disease
AHCVD
Self-reported CHD
0,1 0,5 0.9 1.3 1.7 2.1
Figure 7-1. Risk estimates for the associations of clinical outcomes with long-term exposure to
ambient PM2.5 and PM10
December 2008
7-16
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Zanobetti and Schwartz (2007) studied ICD-coded recurrent MI (ICD-9 410) and post-infarction
CHF (ICD-9 428) among 196,131 Medicare recipients (age > 65 yr; 50% female) discharged alive
following MI hospitalization in 21 cities from 12 U.S. states (1985-1999). In this ecologic, open cohort
study of re-hospitalization among MI survivors (mean duration of follow-up = 3.6 and 3.7 yr for MI and
CHF, respectively), the authors used arithmetic averaging of EPA AQS PMi0 data available in the county
of hospitalization to estimate one-year average exposures. They found that a 10 |ig/m3 increase in PMi0
exposure was associated with 17% (5%, 31%) and 11% (3%, 21%) increases in the hazard of recurrent MI
and post-infarction CHF, respectively. Hazards were somewhat higher among persons aged >75 years.
Hoffman et al. (2006) studied self-reported CHD (MI or revascularization) among 3,399 residents
of Essen and Miilheim, Germany (age range = 45-75 yr; 51% female) at the baseline exam of the Heinz
Nixdorf Recall Study (2000-2003) introduced previously. In this cross-sectional ancillary study, the
authors used dispersion modeling of emissions, climate and topography data to estimate one-year average
exposure to PM2 5 (mean = 23.3 |ig/m3). They found that after adjustment for geographic, demographic
and clinical characteristics, a 10 |ig/m3 increase in exposure was associated with a -45% (-86%, 211%)
decrease in the odds of prevalent CHD.
Rosenlund et al. (2006) studied 2,938 residents of Stockholm County, Sweden (age range = 45-70
yr; 34% female). In this case-control study of 1,085 patients with their first, validated non-fatal MI and an
age-, gender- and catchment-stratified random sample of 1,853 controls without MI (1992-1994), the
authors used street canyon-adjusted dispersion modeling of emissions data to estimate 30-year average
exposure to PM10 (median = 2.4 |ig/m3). They found that the OR for prevalent MI associated with a 10
|ig/m3 increase in PM10was 0.85 (95% CI: 0.50-1.42). The OR for fatal MI was somewhat larger.
In the study of 1030 census enumeration districts in Sheffield, UK described previously,
Maheswaran et al. (2005a) studied 11,407 ICD-10-coded emergency HAs for CHD (120-25) among
199,682 residents (age > 45 yr; 45% female). In this ecologic study, the authors used dispersion modeling
of emissions and climate data to estimate five-year average exposure to PMi0. They found that after
adjusting for smoking prevalence, controlling for socioeconomic factors, and smoothing, the age- and
gender-standardized rate ratios for CHD admission were 1.01 (0.92, 1.11), 1.04 (0.93, 1.15), 0.97 (0.87,
1.08), and 1.07 (0.95, 1.20) across PMi0 quintiles. The linear trend was somewhat stronger for CHD
mortality (see Section 7.3).
Stroke
Two studies examined the long-term PM-stroke association (Maheswaran et al., 2005b; Miller et
al., 2007b). The former examined emergency room HAs in Sheffield, UK using an ecologic design and
December 2008
7-17
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
the latter is based on the prospective cohort study of the WHI OS population (both introduced
previously).
Miller et al. (2007b) found a 10 (.ig/nr1 increase in one-year average PM2 5 exposure was associated
with 28% (95% CI: 2-61), 35% (95% CI: 8-68) and 24% (95% CI: 9-41) increases in the overall hazard of
validated stroke, cerebrovascular disease, and a combined outcome, CVD (MI; revascularization; stroke;
CHD death; cerebrovascular disease), respectively. Hazards were higher within than between cities. For
the combined CVD outcome, they also were also higher among participants at higher than lower quintiles
of body mass index, waist-to-hip ratio, and waist circumference. Counter to the authors' hypothesis, the
PM2 5-CVD association was nonetheless stronger among non-diabetic than diabetic participants.
In the study of 1030 Census of enumeration districts in Sheffield, UK described previously,
Maheswaran et al. (2005b) studied 5,122 ICD-10-coded emergency hospital admissions for stroke (160-
69) among 199,682 residents (age > 45 yr; 45% female) of 1,030 census enumeration districts in
Sheffield, UK (1994-1999). In this ecologic study, the authors used dispersion modeling of emissions and
climate data to estimate five-year average exposure to PMi0. They found that the age- and gender-
standardized rate ratios for stroke admission were 1.05 (95% CI: 0.94-1.17), 1.07 (95% CI: 0.95-1.20),
1.06 (95% CI: 0.94-1.20), and 1.15 (95% CI: 1.01-1.31) across PMi0 quintiles. Linear trend was
somewhat stronger for stroke mortality (see Section 7.6).
Deep Vein Thrombosis
The Italian case-control study (introduced in Section 7.2.1.2) also examined the chronic PM10-DVT
association (Baccarelli et al., 2008). The authors found that a 10 (.ig/nr1 increase in one-year average PM10
exposure was associated with an OR of 1.7 (95% CI: 1.30-2.23) for 70% (30%, 223%) for DVT increase
in the odds of DVT, a finding consistent with the decreases in PT and aPTT also observed among controls
in this context. Strength of the PM10-DVT association was weaker among women and among users of oral
contraceptives, hormone therapy, or either class of endocrinologic therapy.
7.2.7. Overall Summary and Causal Determination
7.2.7.1. PM10
PM2 5 has been the focus of the majority of new research on the long-term effects of exposure to
ambient PM and studies of PMi0 are relatively few. Within this small body of literature, the epidemiologic
evidence is not consistent. Two recent epidemiologic studies of the effect of long-term exposure to
ambient PMi0 report large increases in CVD morbidity from recurrent MI, post-infarction CHF and DVT.
December 2008
7-18
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Two additional studies failed to find meaningful associations with clinical cardiovascular outcomes and
epidemiologic studies of pre-clinical markers of cardiovascular diseases reported largely null findings.
However, two toxicological studies, including one exposing mice to ambient air (including gasses) have
indicated a possible role for PMi0 in the development of atherosclerosis. Methods used in the
toxicological (e.g. inhalation versus intratracheal installation) and epidemiologic (e.g. variable exposure
assessment strategies) were not entirely comparable limiting our ability to fully assess consistency within
and coherence across disciplines. However, in light of two epidemiologic studies showing large
associations, the evidence is determined to be suggestive but not sufficient to infer a causal
relationship.
Atherosclerosis
Several measures of atherosclerosis were examined in the epidemiologic study of the MESA
cohort, including CAC, CIMT, and ABI (Diez Roux et al., 2008). Overall, associations were largely null,
with the exception of a rather weak associations of PMi0 with CIMT. There were no toxicological studies
included in the 2004 PM AQCD that evaluated atherosclerosis development or progression following
subchronic or chronic PMi0 exposure. One new study exposed mice to ambient air (including gases) and
reported decreases in lumen/wall ratio of arteries. A second study demonstrated increased plaque surface
and volume in aortas of rabbits exposed to PMi0 (EHC-93) via intratracheal instillation, as well as
increased monocyte recruitment into these plaques.
Renal and Vascular Function
One epidemiologic study of UACR was conducted (O'Neill et al., 2007). Findings from this study
were largely null with the exception of a weak association of three-year change in log UACR and 20 year
PM10 (O'Neill etal., 2007).
Thromboembolism
Decreases in PT and aPTT were reported in one epidemiologic study of DVT cases and controls in
Lombardy Italy (Baccarelli et al., 2008).
Clinical Outcomes in Epidemiologic Studies
Since 2002, several studies of populations in the U.S. and Europe have examined associations
between CVD morbidity and chronic PM10 exposure. Only Puett et al. (2008), Rosenlund et al. (2006) and
Baccarelli et al. (2008a) examined hospitalization for validated events. Other investigators relied on ICD
codes (Maheswaran et al., 2005a, b; Zanobetti and Schwartz, 2007).
December 2008
7-19
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
With the exception of Puett et al. (2008), a prospective cohort study restricted to women, the
populations under study were > 45 years of age and almost equally balanced in their ratio of men to
women. The studies employed a range of designs including ecologic (Maheswaran et al., 2005a, b), case-
control (Baccarelli et al., 2008; Rosenlund et al., 2006) and open cohort (Zanobetti and Schwartz, 2007).
Definitions of chronic PMi0 exposure were based on either U.S. EPA AQS monitor data (Diez Roux et al.,
2008; O'Neill et al., 2007; Zanobetti and Schwartz, 2007), a combination of monitoring and emissions
data (Puett et al., 2008) or European emissions (Maheswaran et al., 2005a, b; Rosenlund et al., 2006), or
monitor data (Baccarelli et al., 2008). Duration of chronic PMi0 exposure varied across studies.
The studies collectively present somewhat inconsistent an evidence of association between CVD
morbidity and chronic PMi0 exposure. Zanobetti and Schwartz (2007) and Baccarelli et al. (2008a) found
large increases in the adjusted risk of recurrent MI, post-infarction CHF, and DVT with standardized
increments in PMi0. Puett et al. (2008) reported large increases in the risk of CHD death but not incident
CHD or non-fatal MI. Maheswaran et al. (2005b) and Rosenlund et al. (2006) did not find association
between PMi0 and risk of hospitalization for stroke, CHD, or MI. The striking evidence for effect
modification of the PMi0-DVT association by gender and endocrinologic therapy presented by Baccarelli
et al. (2008a) has not been observed again within this body of literature.
7.2.7.2. PM10-2.5
No epidemiologic or toxicological studies of long-term exposure to ambient PM10-2.5 have been
conducted to date. Evidence is inadequate to infer the presence or absence of a causal relationship.
7.2.7.3. PM2.5
Epidemiologic evidence of the adverse effect of PM2 5 on subclinical markers of atherosclerosis is
available from the majority of recent studies on this topic. In addition, a large U.S. study reports
associations of 1-year average PM2 5 concentration with cardiovascular diseases among post-menopausal
women. Further, modification of the PM2 5-CVD association by smoking status and use of anti-
hyperlipidemics has been reported in more than one epidemiologic study. The toxicological studies
provide evidence for accelerated development of atherosclerosis in ApoE" " mice exposed to CAPs for
4-6 months in Tuxedo, NY. Another CAPs study conduced in southern California demonstrated increased
lesion area similar to that observed, with Tuxedo, NY CAPs and the effect was attributable to ultrafine
traffic PM that contained particles in the fine size range (0.18 |im). Two additional toxicological studies of
CAPs from Tuxedo, NY showed effects on coagulation, experimentally-induced hypertension, and
vascular reactivity. The two available studies of clinical cardiovascular disease outcomes did not report
December 2008
7-20
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
consistent results. Still, evidence of short-term PM2 5 effects (e.g. increased mortality, incidence and
progression of cardiovascular disease) reported in a large number of studies spanning multiple disciplines,
supports a role for long-term exposure to PM2 5 in cardiovascular disease. Based on the above findings,
the epidemiologic and toxicological evidence is sufficient to infer a relationship that is likely to be
causal between long-term PM2.5 exposures and cardiovascular morbidity.
Atherosclerosis
Several epidemiologic analyses have reported associations between PM2 5 exposure and subclinical
measures of atherosclerosis including CAC (Diez Roux et al., 2008; Hoffmann et al., 2007), CIMT (Diez
Roux et al., 2008; Kunzli et al., 2005), AAC (Allen et al., in press) and ABI (Diez Roux et al., 2008).
Findings from these studies are consistent with regard to the observed effects modification. There is
consistency in the findings of these studies. Two studies report larger increases in the long-term
PM2 5-CIMT and PM2 5-AAC associations among users than non-users of anti-hyperlipidemics (Allen et
al., in press; Kunzli et al., 2005). Similarly, the long-term PM2 5-CIMT and PMio-CHD death associations
are stronger in never compared to former or current smokers (Kunzli et al., 2005; Puett et al., 2008).
Nonetheless, the remaining associations in MESA, including those of baseline CAC and ABI with 20-
year mean PM2 5, were largely null (Diez Roux et al., 2008).
A group of new toxicological studies demonstrate increased plaque and lesion areas, lipid
deposition and content, and TF in aortas of ApoE_/~ mice exposed to CAPs for 4-6 months. An additional
study of traffic-derived CAPs demonstrated increased atherosclerotic lesion area in the aorta that was
greater for the ultrafine PM, although the upper size fraction included PM in the fine size (0.18 |im).
Systemic Inflammation and Blood Coagulation Markers
One CAPs study demonstrated elevated TF expression in aortas of ApoE~~mice. Three new
toxicological studies utilized diesel or gasoline exhaust, or HWS over a 6-month exposure period and the
findings were largely inconsistent. The limited effects reported in the latter studies may be attributable to
the gases.
Renal and Vascular Function
Cross-sectional and longitudinal epidemiologic analyses of PM2 5 and UACR in the MESA cohort
revealed no evidence of an effect (O'Neill et al., 2007). Auchincloss et al. (2008) reported increased blood
pressure with 60-day average PM2 5 concentration. A toxicological study of 10 weeks exposure duration
did not show changes in MAP with CAPs, but indicated a CAPs-related potentiation of experimentally-
induced hypertension and altered vasoreactivity.
December 2008
7-21
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Autonomic Function
A recent toxicological study reported chronic decreases in HR over 4-5 month CAPs exposure in
ApoE"" mice, with biphasic responses in HRV (SDNN and rMSSD) observed.
Ciinicai Outcomes in Epidemiologic Studies
Two epidemiologic studies of the PM2 5-CVD morbidity relationship focused on clinical CVD
events: in one case, on incident, validated MI, coronary revascularization, and stroke in 36 U.S.
metropolitan areas (Miller et al., 2007b), and in the other, on prevalent, self-reported CHD in Essen and
Miilheim, Germany (Hoffmann et al., 2006). Miller et al. (2007b) was a prospective, cohort study with the
population restricted to women (Miller et al., 2007b). Authors used arithmetic averaging of year 2000
AQS PM2 5 data at the monitor most proximate to each participant's geocoded U.S. Postal Service ZIP
code. The one year average PM2 5 exposure used in the German study was based on dispersion-modeled
emissions data (Hoffmann et al., 2006). The inconsistent findings between theswe two studies may be
driven by differences in study design and location. Miller et al. (2007b) found large increases in the
adjusted risk of MI, revascularization, and stroke with standardized increments in PM25, but for the same
increment, Hoffman et al. (2006) found no such increase in the odds of prevalent CHD. Furthermore,
striking evidence for effect modification by anthropometric measures (e.g., BMI and waist-to-hip ratio)
presented by Miller et al. (2007b) has been tested, e.g. by Diez Roux et al. (2008), but not observed again
within this body of literature.
7.2.7.4. Ultrafine PM
Only one toxicological study of long-term exposure to ulftrafine PM has been conducted to date.
Evidence from one study alone is inadequate to infer the presence or absence of a causal
relationship. Increased plaque size was found in ApoE" " mice exposed for 40 days to ultrafine PM
derived from traffic that included particles in the fine size range (0.18 |im): effects were also observed in
high density lipoprotein and lipid peroxidation was elevated in the liver. Another study evaluated the
effects of 50-day exposures of whole and filtered gasoline exhaust in ApoE"" mice and reported increases
in gene products involved in atheromatous plaque formation and/or degradation, but it appeared that these
effects were due to the gaseous emissions.
December 2008
7-22
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Table 7-2. Characterization of ambient PM concentrations from studies of clinical cardiovascular
diseases.
Reference
Location
Mean Annual
Concentration (|jg/m3)
Upper Percentile
Concentrations (|jg/m3)
PMw
Zanobetti and Schwartz
(2007)
21 U.S. Cities
28.8
Overall range NR
Rosenlund et al. (2006)
Stockholm, Sweden
30 y avg PM10 (traffic)
Cases: 2.6
Controls: 2.4
5th-95th Percentile
0.5-6.0
0.6-5.9
Maheswaran (2005a)
Sheffield, UK
Range of means in each quintile: 16-23.3
NR
Baccarelli et al. (2007b)
Lombardy, Italy
Sep-Nov: 51.2
Dec-Feb: 68.5
Mar-May:64.1
Jun-Aug: 44.3
148.9
238.3
158.5
94.7
PMis
Miller et al. (2007b)
WHI: 36 Metropolitan
areas
Citywide average (year 2000): 13.5
Min-max: 4-19.3
Hoffman et al. (2006)
HNRS: 2 Cities Germany
23.3
NR
WHI: Womens Health Initiative
HNRS: Hans Nixdorf Recall Study
7.3. Respiratory Effects
Several cohort studies reviewed in the 2004 PM AQCD provided evidence for relationships
between long-term PM exposure and effects on the respiratory system. In 12 southern California
communities in the Children's Health Study (CHS), Gauderman et al. (2000; 2002) found that the largest
decreases in lung function growth among school children were associated with long-term exposure to PM
Declines in pulmonary function were reported with all three major PM size classes - PMi0, PMi0-2.5 and
PM2 5- though the three PM measures were highly correlated. In an earlier cross-sectional analysis, Peters
et al. (1999) found no significant relationships between respiratory symptoms and long-term exposure to
PM10. McConnell et al. (1999), in another analysis of data from the CHS cohort, reported an increased
risk of bronchitis symptoms in children living in communities with higher PM concentrations. These
results were found to be consistent with results of cross-sectional analyses of the 24-cities study by
Dockery et al. (1996) and Raizenne et al. (1996), that had been assessed in the 1996 PM AQCD. These
studies reported associations between decreased peak flow and increased bronchitis rates with fine
particle sulfate and acidity. However, the high correlation of PMi0 and acid precluded clear attribution of
the bronchitis effects reported by McConnell et al. (1999) to PM specifically. Finally, among a subset of
December 2008
7-23
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
children in the CHS (n = 110) who moved to other locations during the study period, Avol et al. (2001)
reported that those subjects who moved to areas of lower PM10 showed increased growth in lung function
compared with subjects who moved to communities with higher PM10 concentrations.
The 2004 PM AQCD (U.S. EPA, 2004) concluded that the evidence for an association between
long-term exposure to PM and respiratory effects was inconsistent and potentially confounded by high
correlations between studied pollutants. Quantitatively, Gauderman et al. (2002) reported declines for
FEVi for PMio and PM2 5 of-0.04 mL (95% CI: -0.2 to 0.12) and -0.17 mL (95% CI: -0.47 to 0.13), per
10 |ig/m3 increase, respectively; and McConnell et al. (1999) reported increased ORs for bronchitic
symptoms in asthmatics for PMi0 and PM25 of 1.19 (95% CI: 1.05-1.36) and 1.25 (95% CI: 0.93-1.74)
per 10 (ig/m3 increase, respectively. Very few subchronic and chronic toxicological studies investigating
respiratory effects were available in the 2004 PM AQCD. However, the 2002 EPA Health Assessment
Document for DE reported that long-term exposure to DE was associated with histopathology including
alveolar histiocytosis, aggregation of alveolar macrophages, tissue inflammation, increased
polymorphonuclear leukocytes, hyperplasia of bronchiolar and Type 2 epithelial cells, thickened alveolar
septa, edema, fibrosis, emphysema and lesions of the trachea and bronchi. Since then a number of
inhalation studies have been conducted using CAPs, urban air, DE and woodsmoke.
Recent epidemiologic literature focuses on prospective cohort studies, which found consistent,
positive associations between long-term exposure to PM and PM2 5 and respiratory morbidity. Subchronic
and chronic toxicological studies provide some evidence of altered pulmonary function, amid
inflammatory oxidative responses and histopathological changes following PM2 5 exposures. These results
are summarized below; further details of these studies are summarized in Annexes D and E.
7.3.1. Respiratory Symptoms and Disease Incidence
7.3.1.1. Epidemiologic Studies
New longitudinal cohort studies provide the best evidence to evaluate the relationship between
long-term exposure to ambient PM and increased incidence of respiratory symptoms or disease. A
summary of the mean PM concentrations reported for the studies characterized in this section is presented
in Table 7-3.
Bayer-Oglesby et al. (2005) examined the decline of ambient pollution levels and improved
respiratory health demonstrated by a reduction in respiratory symptoms and diseases in school children
(n = 9591) in Switzerland. They state that if reduced air pollution exposure resulted in improved
respiratory health of children, this would argue in favor of a causal relation. Further, the average reduction
December 2008
7-24
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
of symptom prevalence would be more pronounced in areas with stronger reduction of air pollution
levels. The average decline of PM10 between 1993 and 2000 across the nine study regions was 9.8 (ig/m3
(29%). Roosli et al. (2000; 2001) have demonstrated that PM10 levels are homogeneously distributed
within regions of Switzerland and are not significantly affected by local traffic, justifying the single-
monitor approach for assignment of PMi0 exposures. Declining levels of PMi0 were associated with
declining prevalence of chronic cough, bronchitis, common cold, nocturnal dry cough, and conjunctivitis
symptoms, but no significant associations were reported for wheezing, sneezing, asthma, and hay fever, as
shown in Figure 7-2. In Figure 7-2, Panel (B) illustrates that on an aggregate level across region, the mean
change in adjusted prevalence of chronic cough is associated with the mean change in PMi0 levels
(r = 0.78; p = 0.02). Similar associations were seen for nocturnal dry cough and conjunctivitis symptoms
and PMio levels. Gehrig and Buchmann (2003) conducted co-located parallel measurements of PM2 5 and
PMio at seven sites in Switzerland since January, 1998. The long-term averages of the PM2 5/PM10 ratios
of the daily values vary from 0.75 to 0.76. The correlations between daily values of PM25 and PMi0 at all
sites are generally high (0.8 to 0.58 for the seven cities).
December 2008
7-25
DRAFT-DO NOT QUOTE OR CITE
-------
Table 7-3. Characterization of ambient PM concentrations from studies of respiratory
symptoms/disease and long-term exposures.
Reference
Location
Mean Annual
Concentration (|jg/m3>
Upper Percentile
Concentrations (|jg/m3>
PMw
Bayer-Oglesby et al. (2005)
Nine study regions in Switzerland
Max: 46
Dockery et al. (1996)
24 U.S./Canadian communities
23.8
Kim et al. (2004)
San Francisco, CA
30
McConnell et al. (1999)
12 CFIS/CA communities
34.8
Max: 70.7
McConnell et al. (2003)
12 CFIS/CA communities
30.8
Max: 63.5
Pierse et al. (2006)
Leicestershire, UK
1.33
75th: 1.84
PMis
Brauer et al. (2007)
The Netherlands
16.9
75th: 18.1
90th: 19.0
Max: 25.2
Dockery et al. (1996)
24 U.S./Canadian communities
14.5
Islam et al. (2007)
12 CFIS/CA communities
Max: 29.5
Kim et al. (2004)
San Francisco, CA
12
McConnell et al. (1999)
12 CFIS/CA communities
15.3
Max: 31.5
McConnell et al. (2003)
12 CFIS/CA communities
13.8
Max: 28.5
December 2008
7-26
DRAFT-DO NOT QUOTE OR CITE
-------
1.5
1.4
0 1J
1 «
•S „
1 11
1 1.0
* «
E
1 0-8
0 0.7
1 116
g 0.5
0.4
0.3
5"
< 0.2
0.1
0.0
/ y / / f y /> S s
*.
a O
*Bi
-20 -15 -10 -5
Mean change of annual
average PM10 (pata3)
1 Adjusted for age, sex, nationality, parental education, number of siblings; farming status, low birth weight, breastfeeding, child who smokes, family history of asthma, bronchitis, and/or
atopy, mother who smokes, indoor humidity, mode of heating and cooking, carpeting, pets allowed in bedroom, removal of carpet and/or pets for health reasons, person who completed
questionnaire, month when questionnaire was completed, number of days with the maximum temperature <0°C, and belief of mother that there is an association between environmental
exposures and children's respiratory health.
2 An: Anieres. Be: Bern. Bi: Biel. Ge: Geneva. La:, Langnau. Lu: Lugano. Mo: Montana. Pa: Payerne. Zh: Zurich.
1 A matched case-control study of infant bronchiolitis (ICD 9 code 466.1) hospitalization and two
2 measures of long-term exposure - the month prior to hospitalization (subchronic) and the lifetime average
3 (chronic) - to PM2 5 and gaseous air pollutants in the South Coast Air Basin of southern California was
4 conducted by Karr et al. (2007) in 18,595 infants born between 1995-2000. For each case, 10 controls
5 matched on date were randomly selected from birth records. Exposure was based on PM2 5 measurements
6 collected every third day; the mean distance between the subjects' residential ZIP code and the assigned
December 2008
7-27
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
monitor was in the range of 4-6 miles with a maximum distance of 30 miles. For 10-|ig/m3 increases in
both sub-chronic and chronic PM25 exposure an adjusted odds ratio of 1.09 (95% CI: 1.04-1.14) was
observed. In multipollutant model analysis the association with PM2 5 was robust to the inclusion of
gaseous pollutants. Also, Morgenstern et al. (2008) provides a preliminary examination of modeled PM
data at birth addresses and 1-year and 2-year incidence of respiratory symptoms that show initial positive
findings.
McConnell et al. (2003) conducted a prospective study examining the association between air
pollution and bronchitic symptoms in 475 children with asthma in 12 Southern California communities as
part of the CHS from 1996 to 1999. They investigated both of the differences between communities with
4-year average and yearly variation in pollutants (including PMi0, PM2 5, PMi0.2.5, EC, and OC) within-
communities. Based on a 10 |ig/m3 change in PM25, within-communities effects were larger (OR 1.90
[95% CI: 1.10-2.70]) than those for between-communities assessment (OR 1.30 [95% CI: 1.10-1.50]).
The OR for the 10 |ig/m3 range in 4-yr avg PM2 5 concentrations across the 12 communities was 1.29
(95% CI: 1.06-1.58). Similar results were reported for PMi0 and PM10-2.5 but the effect estimates were
smaller in magnitude and generally not statistically significant. Within-community associations were not
confounded by any time-fixed personal covariates in this study. In two pollutant models, the within-
community effect estimates for PM2 5 and OC were significant in the presence of several other pollutants.
The single-pollutant effect of PM25 (B = 0.085/(ig/m3) was only modestly attenuated by other pollutants
and remained significant after adjusting for other pollutants. The effects of PM25 were markedly reduced
after adjusting for N02 or OC. The between-community effect estimates were generally not significant in
the presence of other pollutants in two-pollutant models.
Pierse et al. (2006) studied the association between primary PMi0 (particles directly emitted from
local sources/traffic) and the prevalence and incidence of respiratory symptoms in a randomly sampled
cohort of 4400 children (aged 1-5 years) in Leicestershire, England surveyed in 1998 and again in 2001.
Annual exposure to primary PMi0 was calculated for the home address using the Airviro statistical
dispersion model. After adjusting for confounders, mean annual exposure to locally generated PMi0 was
associated with an increased prevalence of cough without a cold in both the 1998 and 2001 surveys: 1998
OR 1.21 (95% CI: 1.07-1.38), n = 2164; 2001 OR 1.56 (95% CI: 1.32-1.84), n = 1756.
Kim et al. (2004) conducted a school-based cross-sectional study in the San Francisco metropolitan
area in 2001 comprised of 10 neighborhoods to examine the relationship between traffic-related pollutants
and current bronchitic symptoms and asthma obtained by parental questionnaire (n = 1109). They found
associations (per 10 |ig/m3 increase in PM) between traffic-related pollutants and bronchitic (PMi0 1.02
[95% CI: 0.99-1.05]; PM2 5 1.33 [95% CI: 1.00-2.01) and asthma (PM10 OR= 1.01 [95% CI: 0.97-1.06;
PM2 5 OR = 1.00 95% CI 0.65 to 1.75) symptoms in the past 12 months.
December 2008
7-28
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
In the CHS discussed earlier, Islam et al. (2007) examined the hypothesis that ambient air pollution
attenuates the reduced risk for childhood asthma that is associated with higher lung function (n = 2057).
At each age a distribution of pulmonary functions exists. Haland et al. (2006) found evidence that
children with high lung function have a reduced risk for asthma. Islam et al. (2007) hypothesize that
evolutionary selection has resulted in lung function characteristics that promote better respiratory health.
Therefore, better lung function may be a marker for lower susceptibility to airway pathophysiology. Islam
et al. (2007) used the CHS data to study how the association of asthma incidence with lung function is
modified by long-term PM exposure. They used linear regression to adjust the log of sex-specific lung
function for various known covariates. From that regression, they obtained the percent predicted lung
function (PPFL) variable. This was 100 times the antilog of the residuals. Next, for each of 12
communities, the Cox proportional hazards model was used to relate baseline PPFL (year-1994) and other
individual level covariates such as race/ethnicity to the incidence of asthma during the subsequent years
up to 2003. In that regression, the PPFL was scaled by dividing by its range from the 10th to the 90th
percentile. The estimated 12 lung function coefficients were then used as the dependent variable in a
meta-regression on the 12 long-term average community-specific pollution levels. The hazard ratio was
also calculated for 3 categories of PPFL to better explain the relationship.
The incidence rate (IR) of newly diagnosed asthma increased from 9.5/1000 person-years for
children with percent-predicted FEF25.75 values > 120% to 20.4/1000 person-years for children with
FEF25.75 value < 100%. Over the 10th-90th percentile range for FEF25.75 (57.1), the hazard ratio of new
onset asthma was 0.50 (95% CI: 0.35-0.71). The IR of asthma for FEF25.75 > 120% in the "high" PM2 5
(13.7-29.5 (.ig/ni3) communities was 15.9/1000 person-years compared to 6.4/1000 person-years in "low"
PM2 5 (5.7-8.5 (.ig/ni3) communities. Loss of protection by high lung function against new onset asthma in
the "high" PM2 5 communities was observed for all the lung function measures. Figure 7.3 shows the
effect of PM2 5 on the association of lung function with asthma. Of all the pollutants examined (N02,
PM10, PM2.5, acid vapor, ozone, EC, and OC), PM2 5 appeared to have the strongest modifying effect on
the association between lung function with asthma as it had the highest R2 value (0.42). Over the 10th-
90th percentile range of FEF25.75 (57.1%) the hazard ratio of new onset asthma was 0.34 (95% CI: 0.21-
0.56) in a community with low PM2.5 (less than 13.7 |ig/m3) and 0.76 (95% CI: 0.45-1.26) in a
community with high PM2 5 (equal to or greater than 13.7 |ig/m3). The data do not indicate that PM
exposure increased rates of incident asthma among children with poor lung function at study entry
because rates among those with poor lung function were similar in both low and high pollution
communities.
December 2008
7-29
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
1.00
High PM2 5- Communities
1.50-
~
UP
is ¦—*
x 0.60 -
>
0.30-
¦p 0.90 -
~ 1.20-
Low PM2 5
Communities
o.oo
5
10
15
20
25
30
PM25(Mg/m3)
Source: Islam et al. (2007).
Figure 7-3. Effect of PM2.5 on the association of lung function with asthma. Community-specific
hazard ratio of newly diagnosed asthma over 10-90th percentile range (57.1%) of
FEF25-75% by level of ambient PM2.5 (HQ/m3)- The 12 CHS communities are shown.
During the first four years of life in a birth cohort study (n = 4,000) in The Netherlands, Brauer
et al. (2007) assessed the development of asthma, allergic symptoms, and respiratory infection in relation
to long-term pollution concentration at the home address with a validated model using GIS. PM2 5 was
associated with doctor-diagnosed asthma (OR = 1.32 [95% CI: 1.04-1.69]) for a cumulative lifetime
indicator. Annesi-Maesano et al. (2007) relate individual data on asthma and allergy from 5338 school
children (10.4 ± 0.7 years) attending 108 randomly chosen schools in 6 French cities to the concentration
of PM2.5 (monitored in school yards) that was dichotomized as high (20.7 (.ig/nr1) vs. low (8.7 (.ig/nr1).
Atopic asthma was related to PM2.5 (OR 1.43 [95% CI: 1.07-1.91]). The report is consistent with the
results in an earlier paper (Penard-Morand et al., 2005) in the same sample of children that related the
findings to PM10.
Schikowski et al. (2005) examined the relationship between both long-term air pollution exposure
and living close to busy roads and COPD in the Rhine-Ruhr Basin of Germany from 1985 to 1994 using
consecutive cross-sectional studies. Seven monitoring stations that were not more than 8 km to a woman's
home address provided TSP data from which PM10 was estimated as a conversion factor calculated from
parallel measurement of TSP and PMi0 conducted at the 7 sites in the Ruhr area. Distance to a major road
was determined using GIS. The results of the study suggest that long-term exposure to air pollution from
PM10 and living near a major road might increase the risk of developing COPD and can have a
December 2008
7-30
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
detrimental effect on lung function. All ORs for 5-year exposures were stronger than those for 1-year
exposures.
Goss et al. (2004) conducted a national study examining the relationship between air pollutants and
health effects in a cohort of 11,484 cystic fibrosis (CF) patients over the age of 6 years (mean [SD] 18.4
[10]) enrolled in the Cystic Fibrosis Foundation National Patient Registry in 1999 and 2000. Exposure
was assessed by linking air pollution values from the closest population monitor from the Aerometric
Information Retrieval System (AIRS) with the centroid of the patient's home ZIP code that was within 30
miles. PM2 5 and PMi0 24-h averages were collected every 1 to 12 days. CF diagnosis involves genetic
screening panels. A common severe mutation used is the loss of phenylalanine at the 508th position.
Genotyping was available in 74% of the population and of those genotyped, 66% carried one or more
delta F508 deletions. The mean distance from the patient's zip code to monitors for PMi0 and PM2.5 was
11.5 miles (SD 7.9) and 10.8 miles (SD 7.8) respectfully. After adjusting for confounders a 10 (.ig/ni1 rise
in PMio and PM25 was associated with a 8% (95% CI: 2-15) and 21% (95% CI: 7-33), respectively,
increase in the odds of two or more exacerbations defined as a CF-related pulmonary condition requiring
admission to the hospital or use of home intravenous antibiotics. The estimate for the associations
between pulmonary exacerbations and PMi0 and PM2.5 were attenuated when the models were adjusted
for lung function. Brown et al. (2001) found that particulate deposition was increased in CF and that the
distribution of particle deposition was enhanced in the tracheobronchial regions of poorly ventilated lung
regions in CF patients. Such focal deposition may partially explain the association of particulate air
pollutants and pulmonary exacerbation rate.
7.3.2. Pulmonary Function
7.3.2.1. Epidemiologic Studies
New longitudinal cohort studies have evaluated the relationship between long-term exposure to PM
and changes in measures of pulmonary function (FVC and FEVi, and measures of expiratory flow).
Cross-sectional studies also offer supportive information (see Annex E) and may provide insights derived
from within community analysis. From the major longitudinal cohort studies, associations between
changes in FEVi, FVC, and expiratory flow with a 10 (ig/m3 change in PMi0 standardized per year of
follow up are shown in Figure 7-4. Lung function increases continually through early adulthood with
growth and development, then declines with aging (Stanojevic et al., 2008; Thurlbeck, 1982; Zeman and
Bennett, 2006) thus, the order of results presented in this section is from studies of postnatal exposures
December 2008 7-31 DRAFT-DO NOT QUOTE OR CITE
-------
1 through adulthood. A summary of the mean PM concentrations reported for the studies characterized in
2 this section is presented in Table 7-4.
Study
Location
Age at
baseline
Exposure
duration
Lung
Function
Metric
current
» Indicates direction of Cjl
Oftedal
Oslo
9-10
1 st year
of life
FEF
so
(2008)
only
Gauderman Southern
12004) California
10
8 years
average
mmef
over
8 years
Ofledal
Oslo
9-10
9-10 year
average
current
FEF
50
(2008)
only
Rojas
Martinez
(2007a,b)
Mexico City
8
6 mo
average
growth
over
6 months
-o-
-0-
FEF n
25*75 J
Downs
(2007)
Switzerland
(Never smoked)
41,5 +/
-11.3
11 year
difference
11 year
difference
-0-
FEF
"-0-
1
-20
1 1
-10
i i i i i i
10 -20 -10
1 1 1 1 1 1
10 -100 -50
FEV-j (mL) FVC (mL) Flow Rates (mUs)
Per 10ug/m3PMl0
Figure 7-4. Decrements in FEVi, FVC, FEFso%, FEF25-75, and MMEF and a 10 pg/m3 change in PM10.
Table 7-4. Characterization of ambient PM concentrations from studies of FEVi and long-term
exposures.
Reference
Location
Mean Annual
Concentration (|jg/m3'
Upper Percentile
Concentrations (|jg/m3'
PMw
Downs et al. (2007)
8 cities in Switzerland
9-46
Gauderman et al. (2002)
12 CHS/CA communities
13-78
Gauderman et al. (2004)
12 CHS/CA communities
18-68
Oftedal et al. (2008)
Oslo, Norway
14.5
Raizenne et al. (1996)
22 U.S./Canadian communities
23.8
Max: 32.7
Rojas-Martinez et al. (2007)
Mexico City, Mexico
75.6
75th: 92.2
90th: 112.7
PMis
Dales et al. (2008)
Windsor, Ontario
15.62
95th: 17.17
Gauderman et al. (2002)
12 CHS/CA communities
5-30
Gauderman et al. (2004)
12 CHS/CA communities
6-27
Goss et al. (2004)
U.S.
13.7
75th: 15.9
Gotschi etal. (2008)
21 European cities
3.7-44.7
Oftedal et al. (2008)
Oslo, Norway
12.3
Raizenne et al. (1996)
22 U.S./Canadian communities
14.5
Max: 20.7
December 2008
7-32
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
In a birth cohort (n = 2170) in Oslo, Norway, Oftedal et al. (2008) examined effects of long- and
short-term exposure to PMi0 and PM2 5 on lung function (FVC, FEVi, FEF50%) in early life (first year of
life) and later (9 to 10-year-old children). The EPISODE statistical dispersion model, a GIS approach
(Slordal. 2003), was used for the exposure estimate in which an evaluation concluded that the modeled
PM levels represent the long-term exposure concentrations reasonably well. Only single pollutant models
were evaluated because air pollutants were highly correlated (r = 0.83-0.95). The effects of PM10 in the
first year of life and the results of total lifetime exposure on expiratory flow variables (FEF50o/o) and forced
volumes (FEV, and FVC) are shown in Figure 7.4. No associations were found with shorter exposures
(<30 days) to PMi0 and PM2 5, which suggested permanent impairment. A 10 (.ig/nr1 increase in PMi0 and
PM2 5 was associated with change in adjusted peak respiratory flow of-113.8 mL/s (95% CI: -189.7 to -
39.7) and -161.1 mL/s (95% CI: -261.1 to -58.3), respectively. Adjusting for contextual socioeconomic
factors diminished associations. Results for PM2 5 were similar to those for PMi0. Independent effects of
each pollutant could not be discerned because of their strong correlation related to their traffic-related
primary source emissions. The authors present the notion that the forced volumes results provide
information on central airways, and expiratory flow variables represent peripheral airways.
In an exploratory study, Mortimer et al. (2008) examined the association of prenatal and lifetime
exposure to air pollutants that are most predictive of current pulmonary function in a San Joaquin Valley,
California cohort of 232 children (aged 6-11) with asthma. The data suggested that first and second
trimester PMi0 exposures had a negative effect on pulmonary function at age 6-11 years and may relate to
prenatal exposures affecting the lungs as they begin to develop at 6 weeks gestation continuing through
distinct phases of development.
Nordling et al. (2008) examined the relationship between estimated PM exposure levels and
respiratory health effects in a Swedish birth cohort (n = 4089) of preschool children. Persistent wheezing
(cumulative incidence up to age 4) was associated with exposure to traffic-generated PM10 (OR 2.28
[95% CI: 0.84-6.24] per 10 |ig/m3 increase). Lower peak expiratory flow (not shown in the summary
figure) at age 4 was associated with exposure to traffic-PM10 (-8.93 L/min [95% CI: -17.78 to -0.088]).
The spatial distributions of PM from traffic in the study area were estimated with emission databases and
statistical dispersion modeling. Children were examined at 2 months and 1, 2, and 4 years of age. Using
GIS methods, the average contribution above regional background to the children's residential outdoor air
pollution levels was determined. To evaluate the exposure assessment, the authors compared the estimated
levels of traffic-generated PMi0 with PM2 5 measurements from 42 locations (Hoek et al., 2002).
In a prospective dynamic cohort study consisting of students (n = 3170) who were 8 years of age at
the beginning of the study, who had not been diagnosed with asthma and were located in Mexico City,
December 2008
7-33
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Rojas-Martinez et al. (2007) evaluated the association between long-term exposure to PM10, 03 and N02
and lung function growth every 6 months from April 1996 through May 1999. Exposure data were
provided by 10 air quality monitor stations located within 2 km of each student's school. The
multipollutant model effects of PMi0 over the age of 8 to 10 years of life in this cohort on FVC, FEVi,
and FEF25-75 are shown in Figure 7-4. Single pollutant models showed an association between ambient
pollutants (03, PM10 and N02) and deficits in lung growth. No significant effect of PMi0 was observed on
FEF25-75. While the estimates from two-pollutant models were not substantially different than single
pollutant models, independent effects for pollutants could not be estimated accurately because the traffic-
related pollutants were correlated.
The CHS prospectively examined the relationship between air pollutants and lung function (FVC,
FEVi, MMEF) in a cohort (n = 1759) between the ages of 10 and 18 years, a period of rapid lung
development (Gauderman et al., 2004). Air pollution monitoring stations provided data in each of the 12
study communities from 1994-2000. The results for 03> PM,„ N02, PM2 5, acid vapor, and EC and are
depicted in Figure 7-5. In general, two-pollutant models for any pair of pollutants did not provide a
significantly better fit to the data than the corresponding single-pollutant models due to the strong
correlation between most pollutants. The pollution-related deficits in the average growth in lung function
over the eight-year period resulted in clinically important deficits in attained lung function at the age of
18 years. Since lung development is basically completed by the age of 18 years, it is unlikely that these
clinically significant deficits in lung function will be reversed. Since associations are seen between the
three measures of lung function this suggests that more than one biological process is involved. FVC is
largely a function of the number and size or growth of alveoli. A possible effect of air pollution on lung
development may be airway inflammation, as occurs in bronchiolitis.
December 2008
7-34
DRAFT-DO NOT QUOTE OR CITE
-------
ID-
S'
6-
4-
2-
0
N.O
ON
10-1
8-
"O
n
_u
"O
a
6-
CL
<-£
4-
o
oo
V
2-
LU
0-
Li_
0
R=0.04
P=0,89
~ LB
~ UP
bu
~ ML
~ RV
~ AT
~ SM ~ LM
~ AL
~ LE
~ LA
—I—
35
« LN
i—
55
—I—
65
03 from 10 a.m. to 6 p.m. (ppb)
—i
75
~ UP
R=0.75
P=0.005
~ LM
N02(ppb)
R=0.69
P=0.01
~ RV
~ UP
~ SD
4 6
Acid Vapor (ppb)
10
12
10-
~o
V
5 6-
"O
I 4"
NO
ON
° 2-\
oo '
V
R=U./4
P=0.006
~ UP
~ RV
~ LM
0.0
0.2
0.4 0.6 0.8 1.0
Elemental Carbon (f/g/m3)
1.2
1.4
Source: Adapted from Gauderman et al. (2004).
Figure 7-5. Proportion of 18-year olds with a FEVi below 80% of the predicted value plotted
against the average levels of pollutants from 1994 through 2000 in the 12 southern
California communities of the Children's Health Study.1
1 In Leicester, England, investigators examined the carbon content of airway macrophages in
2 induced sputum in 64 of 114 healthy children 8 to 15 years of age (Grigg et al., 2008; Kulkarni et al.,
3 2006). The carbon content of airway macrophages (Finch et al., 2002; Strom et al., 1990) was used as a
4 marker of individual exposure to PMi0. Near each child's home, modeled exposure to PMi0 was
1 AL = Alpine; AT = Atascadero; LA = Lake Arrowhead; LB = Long Beach; LE = Lake Elsinore; LM = Lompoc; LN = Lancaster; ML = Mira
Loma; RV = Riverside; SD = San Dimas; SM = Santa Maria; LIP = LTpland
December 2008
7-35
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
determined using the Airviro statistical dispersion model (Pierse et al., 2006). The authors reported a
dose-dependent inverse association between the carbon content of airway macrophages and lung function
in children and note that they found no evidence that reduced lung function itself causes an increase in
carbon content. Consistent results were obtained for both FVC and FEF25.75. The authors conclude that
since they directly assessed the carbon content of airway macrophages, their data strengthen the evidence
for a causal association between the inhalation of carbonaceous particles and impaired lung function in
children. HEI Reports (Grigg et al., 2008) suggest caution when interpreting these results: the accuracy of
the estimates on individual PMi0 exposures obtained by using the Airvivo dispersion model was not
validated; potential for confounding by ethnic origin; and concern that the magnitude of the changes in
pulmonary function associated with increased particle area appear large thus casting doubt on the results.
Dales et al. (2008) examined the relationship of pulmonary function and PM measures, other
pollutants, and indicators of motor vehicle emissions in Windsor, Ontario, in a cohort of 2402 school
children, with PM2 5 and PMi0 concentrations estimated for each child's residence at the postal code level.
Each 10 |ig/m3 increase in PM2 5 was associated with a 7.0% decrease in FVC expressed in a percentage
of predicted (p = 0.39).
Some new studies are using individual estimates of exposure to ambient PM to reduce the impact
of exposure error (Downs et al., 2007; Jerrett et al., 2005a). Downs et al. (2007) prospectively examined
9,651 randomly selected adults (18 to 60 years of age) in 8 cities in Switzerland (see also Ackermann-
Liebrich et al., 1997) to ascertain the relationship between reduced exposure to PM10 and age-related
decline in lung function (FVC, FEVi, and FEF25.50). An evaluated statistical dispersion model (Liu et al.,
2007c) provided spatially resolved concentrations of PM10 that enabled assignment to residential
addresses for the participant examinations in 1991 and 2002 that yielded a median decline of 5.3 (ig/m3
(IQR 4.1 - 7.5). The decreasing exposure to PMi0 attenuated the decline in lung function (see Figure 7.4.)
Effects were greater in tests reflecting small airway function. No other pollutant relationships were
evaluated, though a related study indicated that levels of N02 also declined over the same period
(Ackermann-Liebrich et al., 2005). Generalized cross-validation essentially chose a linear fit for the dose-
response curve.
These data show that improvement in air quality may slow the annual rate of decline in lung
function in adulthood indicating positive consequences for public health. Further evidence on
improvement in respiratory health with reduction in air pollution levels was provided in studies conducted
in East Germany related to dramatic emissions reductions after the reunification in 1990 (Fryer and
Collins, 2003; Heinrich et al., 2002; Sugiri et al., 2006). This type of "natural experiment" provides
additional support for epidemiologic findings that relatively low levels of airborne particles have
respiratory effects.
December 2008
7-36
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Figure 7-4 presents risk estimates for three lung function measurements (FEV,. FVC, and flow
rates) from this group of studies. It is important to recognize that these measurements have been made at
different ages in the cohorts of children, and different lung function indicators have been measured, so the
results are not directly comparable. It can be seen, however, that all estimates are negative
(i.e., decreasing lung function) and the pattern of effects are similar between the studies for FVC and
FEVi. Thus, the data are consistent and coherent across several designs, locations, researchers, and other
factors. With cautions noted, the results relating carbon content of airway macrophages to decreased
measures of pulmonary function add plausibility to the epidemiologic findings.
Gotschi et al. (2008) examined the relationship between air pollution and lung function in adults in
the European Community Respiratory Health Survey (ECRHS). FEVi and FVC were assessed at baseline
and after 9 years of follow-up from 21 European centers (followed-up sample n = 5610). PM25 was
measured in 2000-2001 using central monitors. No significant associations were found between city-
specific annual mean PM2 5 and average lung function levels which is in contrast to the results seen by
Ackermann-Liebrich et al. (1997) (SAPALDIA) and Schikowski et al. (2005) (SALIA) which compared
across far more populations than is the case for ECRHS. The authors presented concerns for potential
misclassification and confounding.
As was found in the 2004 PM AQCD, the studies report associations with PM10 and PM2 5, and
most did not evaluate PM10_2.5. Associations have been reported with fine particle components, particularly
EC and OC. Source apportionment methods generally have not been used in these long-term exposure
studies, however, numerous studies have evaluated exposures to PM related to traffic or motor vehicle
emissions. For example, Meng et al. (2007b) investigated the associations between traffic and outdoor
pollution levels and poorly controlled asthma among adults who were respondents to the California
Health Interview Survey and found associations for traffic density and PMi0 but not PM2 5.
7.3.2.2. Toxicological Studies
Urban Air
An important new study evaluated the effects of chronic exposure to ambient levels of urban
particles on lung development in the mouse (Mauad et al., 2008). Both functional and anatomical indices
of lung development were measured. Male and female BALB/c mice were continuously exposed to
ambient or filtered Sao Paolo air for 8 months. Concentrations in the "polluted chamber" vs. "clean
chamber" were 16.8 vs. 2.9 (ig/m3 PM2 5. Thus PM levels were reduced by filtration but not entirely
eliminated. Ambient concentrations of CO, N02 and S02 were 1.7 ppm, 89.4 (ig/m3 and 8.1 (ig/m3,
respectively. Concentrations of gaseous pollutants were assumed to be similar to ambient levels in both
December 2008
7-37
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
chambers. After 4 months, the animals were mated and the offspring were divided into 4 groups to
provide for a prenatal exposure group, a postnatal exposure group, a pre and postnatal exposure group and
a control group. Animals were sacrificed at 15 and 90 days of age for histological analysis of lungs.
Pulmonary pressure-volume measurements were also conducted in the 90 day old offspring. Statistically
significant reductions in inspiratory and expiratory volumes were found in the pre and postnatal exposure
group, but not in the prenatal or postnatal exposure group, compared with controls. These changes in
pulmonary function correlated with anatomical changes which are discussed in Section 7.3.5.1.
Diesel Exhaust
Li et al. (2007) exposed BALB/c and C56BL/6 mice to clean air or to low dose DE (at a PM
concentration of 100 (.ig/nr1) for 7 hours/day and 5 days/week for 1, 4 and 8 weeks. Average gas
concentrations were reported to be 3.5 ppm CO, 2.2 ppm N02, and less than 0.01 ppm S02. AHR was
evaluated by whole body plethysmography at Day 0 and after 1, 4 and 8 weeks of exposure. Short-term
responses are discussed in Section 6.3.11.3. The increased sensitivity of airways to methacholine
(measured as Penh) seen in C57BL/6 but not BALB/c mice at 1 week was also seen at 4 weeks but not at 8
weeks. This study suggests that adaptation occurs during prolonged DE exposure. Influx of inflammatory
cells, histopathology, markers of oxidative stress and effects of antioxidant intervention were also
evaluated (see Sections 7.3.3.2 and 7.3.4.1). Although no attempt was made in this study to determine the
effects of gaseous components of DE on the measured responses, concentrations of gases were very low
suggesting that PM may have been responsible for the observed effects.
Woodsmoke
One study evaluated the effects of subchronic woodsmoke exposure on pulmonary function in
Brown Norway rats, which are considered an animal model of allergy. Rats were exposed 3 h/day and 5
days/week for 4 and 12 weeks to air or to 1000-10,000 (ig/m3 concentrated wood smoke from the pinon
pine which is native to the U.S. Southwest (Tesfaigzi et al., 2002). PM concentrations in the woodsmoke
were 1,000 and 10,000 (ig/m3 PM. The particles in this woodsmoke had a bimodal size distribution with
the smaller size fraction (74%) characterized by a MMAD of 0.405 (.im and the larger size fraction (26%)
characterized by a MMAD of 6.7-11.7 |_im. Many of these larger particles would not be inhalable by the
rat since 8 (.im MMAD particles are about 50% inhalable (Menache et al., 1995). Concentrations of gases
were reported to be 15-106.4 ppm CO, 2.2-18.9 ppm NO, 2.4-19.7 ppm NOx and 3.5-13.8 ppm total
hydrocarbon in these exposures. Respiratory function measured by whole-body plethysmography
demonstrated a statistically significant increase in total pulmonary resistance in rats exposed to
December 2008
7-38
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
1000 |ag/m3 woodsmoke. Additional effects were found at 10,000 (ig/m3. Inflammatory and
histopathological responses were also evaluated (see Sections 7.3.3.2 and 7.3.5.1).
7.3.3. Pulmonary Inflammation
7.3.3.1. Epidemiologic Studies
One epidemiologic study examined the relationship of airway inflammation (eNO) and PM
measures, other pollutants, and indicators of motor vehicle emissions in Windsor, Ontario (Dales et al.,
2008). This cohort of 2402 school children estimated PM2 5 and PMi0 for each child's residence at the
postal code level with an evaluated statistical model (Wheeler et al., 2006b). Each 10 (.ig/nr3 increase in
1 year PM2 5 was associated with a 39% increase in eNO (p = 0.058). Associations between eNO and
PM10-2.5 were positive but not statistically significant.
7.3.3.2. Toxicological Studies
CAPs Studies
An important set of subchronic studies involved exposure of normal (C57BL1/6) mice, ApoE" " and
the doub 1 e-knockout Apo E VLD LR mice to Tuxedo, NY CAPs for 5-6 month (March, April or May
through September 2003; (Lippmann et al., 2005a). The average PM2 5 exposure concentration was
110 |ig/m3. Animals were fed a normal chow diet during the CAPs exposure period. No pulmonary
inflammation was observed in response to CAPs exposure as measured by BALF cell counts and
histology. The lack of a persistent pulmonary response may have been due to adaptation of the lung
following repeated exposures. In fact, a parallel but preliminary study examined CAPs-related gene
expression in the double-knockout animals and found upregulation of numerous genes in lung tissue
(Gunnison and Chen, 2005). A second parallel study found daily variations in CAPs-mediated NFkB
activation in cultured human bronchial epithelial cells, suggesting that transcription factor-mediated gene
upregulation could occur in response to CAPs (Maciejczyk et al., 2005; Maciejczyk and Chen, 2005). It
should be noted that significant cardiovascular effects were observed in these subchronic studies which
are discussed in Section 7.2.1.2.
Araujo et al. (2008) compared the relative impact of ultrafine (0.01-0.18 |im) versus fine
(0.01-2.5 (mi) PM inhalation in ApoE_/" mice following a 40 day exposure (5 hours/day x 3 days/week for
75 total hours). Animals were on a normal chow diet and exposed to PM from November 3-December 12,
2005 in a mobile inhalation laboratory that was parked 300 m from the 110 Freeway in downtown Los
Angeles. Particles were concentrated to -440 |ig/m3 for the fine exposures and -110 (ig/m3 for the
December 2008
7-39
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
ultrafine exposures, representing a roughly 15-fold increase in concentration from ambient levels; the
number concentration of PM in the fine and ultrafine chambers were roughly equivalent (4.56/105 and
5.59/105 particles/cm3, respectively). Over 50% of the ultrafine PM was comprised of organic carbon
compared to only 25% for PM2 5. No major increase in BALF inflammatory cells was found in response
to PM. However ultrafine PM exposure resulted in significant cardiovascular and systemic effects (see
Section 7.2.1).
Diesel Exhaust
Ishihara and Kagawa (2003) exposed Wistar rats to filtered air and DE containing
PMconcentrations of 200, 1,000 and 3,000 (ig/m3 for 16 /day and 6 days/week for 6, 12, 18 or 24 months.
The mass median particle diameter was reported to be between 0.3 and 0.5 |_im. Concentrations of gases
ranged from 2.93-35.67 ppm NOx, 0.23-4.57 ppm S02, 1.8-21.9 ppm CO in the DE exposures.
Statistically significant increases in total numbers of inflammatory cells and neutrophils in BALF were
observed beginning at 6-12 months of exposure to DE containing 1,000 and 3,000 (ig/m3 PM. When rats
were exposed to DE containing 1,000 (ig/m3 PM, which was filtered to remove PM, the inflammatory cell
response was significantly diminished. These results indicate that the PM fraction of DE was mainly
responsible for the observed influx of inflammatory cells into the lung under those exposure conditions.
The PM fraction was also found to mediate the increase in protein levels (see Section 7.3.4.1), decrease in
PGE2 levels and alterations in mucus and surfactant components observed in BALF.
Li et al. (2007) exposed BALB/c and C56BL/6 mice to clean air or to low dose DE (low dose DE
at a PM concentration of 100 (.ig/ni3) for 7 h/day and 5 days/week for 1, 4 and 8 weeks. Average gas
concentrations were reported to be 3.5 ppm CO, 2.2 ppm N02, and less than 0.01 ppm S02. Increases in
numbers of BALF macrophages and total inflammatory cells were observed in BALB/c mice at eight
weeks but not four weeks of DE exposure. Persistent increases in numbers of BALF neutrophils and
Lymphocytes were observed in both strains at four and eight weeks of DE exposure. Persistent increases
in BALF cytokines also were observed at four and eight weeks of DE exposure, although the responses
differed between the two strains. These results should be interpreted with caution since comparisons were
made with Day 0 controls rather than age-matched controls. No histopathological changes in the lungs
were seen at any time point after DE exposure. This study demonstrates differences in pulmonary
responses to low dose DE between 2 mouse strains. Airway hyperresponsiveness, pulmonary
inflammation, markers of oxidative stress and effects of antioxidant intervention were also evaluated (see
Sections 7.3.2.2 and 7.3.4.1). Although no attempt was made in this study to determine the effects of
gaseous components of DE on the measured responses, concentrations of gases were very low suggesting
that PM may have been responsible for the observed effects.
December 2008
7-40
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
In a study by Hiramatsu et al. (2003), BALB/c and C57BL/6 mice were exposed to DE (PM
concentrations 100 and 3000 (.ig/nr1) for 1 or 3 months. Concentrations of gases were reported to be 3.5-
9.5 ppm CO, 2.2-14.8 ppm NOx, and less than 0.01 ppm S02. Modest increases in BALF neutrophils and
lymphocytes were observed in response to DE in both mouse strains at 1 and 3 months. Histological
analysis demonstrated DEP-laden alveolar macrophages in alveoli and peribronchial tissues at both time
points. Bronchus-associated lymphoid tissue developed after 3 months exposure to the higher
concentration of DE in both mouse strains. Mac-1 positive cells (a marker of phagocytic activation of
alveolar macrophages) were also increased in BALF of BALB/c mice exposed to the higher concentration
of DE for 1 and 3 months. Increased expression of several cytokines and decreased expression of iNOS
mRNA was observed in DE-exposed mice at 1 and 3 months. NFkB activation was also noted following 1
month exposure to the lower concentration of DE. No attempt was made in this study to determine the
responses to gaseous components of the DE.
In a study by Reed et al. (2004), healthy Fisher 344 rats and A/J mice were exposed to DE (PM
concentration = 30, 100, 300 and 1000 (ig/m3) by whole body inhalation for 6 h/day, 7 days/week for
either 1 week or 6 months. Concentrations of gases were reported to be 2.0-45.3 ppm NO, 0.1-4.0 ppm
N02, 1.5-29.8 ppm CO and 8-365 ppb S02. Short-term responses are discussed in Section 6.3.3.3 and
6.3.7.1, and sub-chronic systemic effects are presented in Section 7.2.3.1. Six months of exposure resulted
in no measurable effects on pulmonary inflammation. However numerous black particles were observed
within alveolar macrophages after 6 months of exposure.
Seagrave et al. (2005b) evaluated pulmonary responses in male and female CDF (F-344)/CrlBR
rats exposed 6 h/day for 6 months to filtered air or DE at concentrations ranging from 30-1000 (ig/m3 PM.
Concentrations of gases were reported for the highest exposure as 45.3 ppm NO, 4.0 ppm N02, 29.8 ppm
CO and 2.2 ppm total vapor hydrocarbon. No changes in BALF cells were noted. A small decrease in
TNFa was seen in BALF of female rats exposed to the highest concentration of DE for 6 months.
Pulmonary injury also was evaluated (Section 7.3.5.1). Thus changes in BALF markers are modest and
gender specific.
Woodsmoke
Seagrave et al. (2005b) also evaluated pulmonary responses in male and female CDF (F344)/CrlBR
rats exposed 6 h/day for 6 months to filtered air or HWS concentrations ranging from 30-1000 (ig/m3 PM.
Concentrations of gases were reported for the highest exposure as 3.0 ppm CO and 3.1 ppm total vapor
hydrocarbon. A small increase in BALF neutrophils was observed in male rats exposed to the lowest
concentration of HWS. Pulmonary injury was evaluated in Section 7.3.5.1. Female rats exhibited a
decrease in BALF MIP-2 at the highest concentration of HWS. Pulmonary injury also was evaluated
December 2008
7-41
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
(Section 7.3.5.1). In general, responses to HWS were more remarkable than responses to DE seen in the
same study. However these gender-specific responses are modest and difficult to interpret.
In a study by Reed et al. (2006), Fisher 344 rats, SHR rats, A/J mice and C57BL/6 mice were
exposed to clean air or HWS (PM concentrations 30, 100, 300 and 1000 |_ig/m3' by whole body inhalation
for 6 h/day, 7 days/week for either 1 week or 6 months. Concentrations of gases ranged from 229.0-
14887.6 mg/m3 for CO, 54.9-139.3 |_ig/m3 for ammonia, and 177.6- 3455.0 |_ig/m3 nonmethane VOC in
these exposures. Short-term responses are discussed in Section 6.3.7.1 and sub-chronic effects are
presented in Section 7.2.3.1. Histological analysis of lung tissue showed minimal increases in alveolar
macrophages. The effects of HWS on bacterial clearance are discussed below (Section 7.3.7.1).
Another study evaluated the effects of subchronic woodsmoke exposure in Brown Norway rats,
which are considered an animal model of allergy. Rats were exposed 3 h/day and 5 days/week for 4 and
12 weeks to air or to concentrated woodsmoke from the pinon pine which is native to the U.S. Southwest
(Tesfaigzi et al., 2002). PM concentrations in the woodsmoke exposures were 1,000 and 10,000 (ig/m3
The particles in this woodsmoke had a bimodal size distribution with the smaller size fraction (74%)
characterized by a MMAD of 0.405 (.un and the larger size fraction (26%) characterized by a MMAD of
6.7-11.7 |_im. Many of these larger particles would not be inhalable by the rat since 8 |_im MMAD particles
are about 50% inhalable (Menache et al., 1995). Concentrations of gases were reported to be 15-106.4
ppm CO, 2.2-18.9 ppm NO, 2.4-19.7 ppm NOx and 3.5-13.8 ppm total hydrocarbon in these exposures.
Numbers of alveolar macrophages in BALF were significantly increased in rats exposed to 1000 (ig/m3
woodsmoke for 12 weeks, but no changes were seen in numbers of other inflammatory cells. A large
percent of BALF macrophages contained carbonaceous material. Histological analysis of lung tissue
showed minimal to mild inflammation in the epiglottis of the larynx in rats exposed to both
concentrations of woodsmoke.
7.3.4. Pulmonary Oxidative Response
7.3.4.1. Toxicological Studies
Diesel Exhaust
Li et al. (2007) exposed BALB/c and C56BL/6 mice to clean air or to low dose DE (PM
concentration 100 (.ig/m3) for 7 h/day and 5 days/week for 1, 4 and 8 weeks. Average gas concentrations
were reported to be 3.5 ppm CO, 2.2 ppm N02, and less than 0.01 ppm S02. Markers of oxidative stress
and effects of antioxidant intervention were evaluated in this model. While HO-1 mRNA and protein were
increased in lung tissues of both mouse strains after 1 week of DE exposure (see Section 6.3.4.2), at
December 2008
7-42
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
8 weeks of DE exposure, HO-1 protein levels remained high in C57BL/6 mice but returned to control
values in BALB/c mice. This study demonstrates differences in pulmonary responses to low dose DE
between two mouse strains. Furthermore, this study suggests that adaptation occurs in BALB/C mice
during prolonged DE exposure since the increase in HO-1 protein seen in both strains at 1 week of
exposure was only seen in C57BL/6 mice at 8 weeks. Airway hyperresponsiveness (Section 7.3.2.2) and
pulmonary inflammation (Section 7.3.3.2) were also evaluated. Although no attempt was made in this
study to determine the effects of gaseous components of DE on the measured responses, concentrations of
gases were very low. This suggests that PM may have been responsible for the observed effects.
7.3.5. Pulmonary Injury
7.3.5.1. Toxicological Studies
Urban Air
Prolonged exposure to low levels of ambient air pollution beginning in early life has been linked to
secretory changes in the nasal cavity of mice, specifically increased production of acidic mucosubstances
(Pires-Neto et al., 2006). Six day-old Swiss mice were continuously chamber exposed to ambient or
filtered Sao Paulo air for 5 months. Concentrations in the "polluted chamber" vs. "clean chamber" were
(in (ig/m3) 59.52 vs. 37.08 for N02, 12.52 vs. OforBC, and 46.49 vs. 18.62 for PM2 5. Thus, pollutant
levels were reduced by filtration but not entirely eliminated. Compared to filtered air, exposure to ambient
air resulted in increased total mucus and acidic mucus in the epithelium lining the nasal septum, but no
statistically significant differences in other parameters (amount of neutral mucus, volume proportions of
neutral mucus, total mucus, or nonsecretory epithelium, epithelial thickness, or ratio between neutral and
acidic mucus). The physicochemical properties of mucus glycoproteins are critical to the protective
function of the airway mucus layer. Acidified mucus is more viscous, and is associated with a decrease in
mucociliary transport. Thus acidic mucosubstances may represent impaired defense mechanisms in the
respiratory tract.
An important new study evaluated the effects of chronic exposure to ambient levels of urban
particles on lung development in the mouse (Mauad et al., 2008). Both functional and anatomical indices
of lung development were measured. Male and female BALB/c mice were continuously exposed to
ambient or filtered Sao Paulo air for 8 months. Concentrations in the "polluted chamber" vs. "clean
chamber" were 16.8 vs. 2.9 (ig/m3 PM2 5. Thus PM levels were reduced by filtration but not entirely
eliminated. Ambient concentrations of CO, N02 and S02 were 1.7 ppm, 89.4 (ig/m3 and 8.1 (ig/m3,
respectively. Concentrations of gaseous pollutants were assumed to be similar to ambient levels in both
December 2008
7-43
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
chambers. After 4 months, the animals were mated and the offspring were divided into 4 groups to
provide for a prenatal exposure group, a postnatal exposure group, a pre- and postnatal exposure group
and a control group. Animals were sacrificed at 15 and 90 days of age for histological analysis of lungs.
Histological analysis demonstrated the presence of mild foci of macrophages containing black dots of
carbon pigment in the prenatal and postnatal exposure group at 90 days. In addition, the alveolar spaces of
15-day old mice in the prenatal and postnatal exposure group were enlarged compared with controls.
Morphometric analysis demonstrated statistically significant decreases in surface to volume ratio at 15
and 90 days in the pre and postnatal group compared with controls. Since alveolarization is normally
complete by 15 days of age, these results suggest incomplete alveolarization in the 15-day old group and
an enlargement of air spaces in the 90-day old group. These anatomical changes correlated with
decrements in pulmonary function which are discussed in Section 7.3.2.2.
Kato et al. (2003) exposed Wistar rats to roadside air contaminated mainly with automobile
emissions (55.7 to 65.2 ppb N02 and 63 to 65 (ig/m3 suspended PM [SPM]) and examined the effects on
respiratory tissue after 24, 48, or 60 weeks of exposure. The surface of the lungs was light gray in color
after all durations of exposure, and BC particle deposits accumulated with prolonged exposure. These
characteristics were not evident in filtered air-exposed control animals, although filtered air contained low
levels of air pollutants (< 6.2 ppb N02 and 15 (ig/m3 SPM). The most common change observed using
transmission electron microscopy was the presence of particle laden (anthracotic) alveolar macrophages,
or anthracosis, in a wide range of pulmonary tissues, including the submucosa, tracheal- and bronchiole-
associated lymph nodes, alveolar wall and space, pleura, and perivascular connective tissue. These
changes were evident after 24 weeks and increased with duration of exposure. Other changes included
increases in the number of mucus granules in goblet cells, mast cell infiltration (but no degranulation)
after 24 weeks, and increased lysosomes in ciliated cells, some altered morphology of Clara cells, and
hypertrophy of the alveolar walls after 48 weeks. No goblet cell proliferation was observed, but slight,
variable acidification of mucus granules appeared after 24 and 48 weeks and disappeared after 60 weeks.
Anthracotic macrophages were seen in contact with plasma cells and lymphocytes in the lymphoid tissue,
suggesting immune cell interaction in the immediate vicinity of particles. Even after 60 weeks, no lymph
node anthracosis was observed in the filtered air group.
In a post-mortem study of lung tissues from 20 female lifelong residents of Mexico City, a high PM
locale, histology demonstrated significantly greater amounts of fibrous tissue and muscle in the airway
walls compared to subjects from Vancouver (Churg et al., 2003), a city with relatively low PM levels.
Electron microscopy showed carbonaceous aggregates of ultrafine particles, which the authors conclude
penetrate into and are retained in the walls of small airways. The study shows an association between
retained particles and airway remodeling in the form of excess muscle and fibrotic walls. The subjects
December 2008
7-44
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
were deemed suitable for examination based on never-smoker status, no use of biomass fuels for cooking,
no known occupational particle/dust exposure, death by cause other than respiratory disease, and
extended residence in each locale (lifelong for Mexico City and >20 years for Vancouver). However,
subjects from the two locales were not matched with respect to ethnicity, sex (20 females from Mexico
City vs. 13 females and 7 males from Vancouver), or mean age at death (66 ± 9 vs. 76 ± 11), and other
possibly influential factors such as exercise or diet were not considered.
Diesel Exhaust
Ishihara and Kagawa (2003) exposed Wistar rats to filtered air and DE (PM concentrations of 200,
1000 and 3000 (ig/m3) for 16/day and 6 days/week for 6, 12, 18 or 24 months. Concentrations of gases
ranged from 2.93-35.67 ppm NOx, 0.23-4.57 ppm S02, 1.8-21.9 ppm CO in the DE exposures. A
statistically significant increase in BALF protein was observed at 12 months of exposure to DE
containing 1000 (ig/m3 PM. This response was attenuated when the DE was filtered to remove PM.
Pulmonary inflammation was noted also (Section 7.3.4).
Seagrave et al. (2005b) evaluated pulmonary responses in male and female CDF (F-344)/CrlBR
rats exposed 6 h/day for 6 months to filtered air or DE at PM concentrations ranging from 30-1000 (ig/m3.
Concentrations of gases were reported for the highest exposure as 45.3 ppm NO, 4.0 ppm N02, 29.8 ppm
CO and 2.2 ppm total vapor hydrocarbon. A small increase in LDH was seen in BALF of female rats
exposed to the highest concentration of DE for 6 months. Pulmonary inflammation was evaluated also
(Section 7.3.3.2). The changes in BALF markers in this study were modest and gender specific.
Woodsmoke
Seagrave et al. (2005b) also evaluated pulmonary responses in male and female CDF
(F-344)/CrlBR rats exposed 6 h/day for 6 months to filtered air or HWS with PM concentrations ranging
from 30-1000 (ig/m3. Concentrations of gases were reported for the highest exposure as 3.0 ppm CO and
3.1 ppm total vapor hydrocarbon. Increases in BALF LDH and protein were seen in male but not female
rats exposed to 100 and 300 (ig/m3 HWS. Female rats exhibited a decrease in BALF glutathione at the
highest concentration of HWS. Decreases in BALF alkaline phosphatase were found in both males and
females exposed to 1000 (ig/m3 HWS. Male rats exposed to 100 and 300 |_ig/m3 HWS exhibited a
decrease in BALF |3-glucuronidase activity. Pulmonary inflammation was evaluated also (Section
7.3.3.2). These changes in BALF markers in this study were modest and gender specific.
Another study evaluated the effects of subchronic woodsmoke exposure in Brown Norway rats.
Animals were exposed 3 h/day and 5 days/week for 4 and 12 weeks to air or concentrated wood smoke
from the pinon pine which is native to the U.S. Southwest (Tesfaigzi et al., 2002). PM concentrations in
December 2008
7-45
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
the woodsmoke exposure were 1,000 and 10,000 (ig/m3 The particles in this woodsmoke had a bimodal
size distribution with the smaller size fraction (74%) characterized by a MMAD of 0.405 (.un and the
larger size fraction (26%) characterized by a MMAD of 6.7-11.7 |_im. Many of these larger particles
would not be inhalable by the rat since 8 (.iM MMAD particles are about 50% inhalable (Menache et al.,
1995). Concentrations of gases were reported to be 15-106.4 ppm CO, 2.2-18.9 ppm NO, 2.4-19.7 ppm
NOx and 3.5-13.8 ppm total hydrocarbon in these exposures. Exposure to 1,000 |_ig/m3 woodsmoke for
12 weeks resulted in a statistically significant increase in Alcian Blue- (AB) and Periodic Acid Schiff-
(PAS) positive airway epithelial cells compared to controls, indicating an increase in mucous secretory
cells containing neutral and acid mucus, respectively. More significant histopathological responses were
found following exposure to 10,000 (ig/m3 of DE. Pulmonary function and inflammation were evaluated
also but are not discussed here due to the extremely high exposure level (Sections 7.3.2.2 and 7.3.3.2).
7.3.6. Allergic Responses
7.3.6.1. Toxicological Studies
Diesel Exhaust
Exposure to relatively low doses of DE has been shown to exacerbate asthmatic responses in OVA
sensitized and challenged BALB/c mice (Matsumoto et al., 2006). Mice were intraperitoneally sensitized
and intranasally challenged lday prior to inhalation exposure to DE (PM concentration 100 (ig/m3; CO,
3.5 ppm; N02, 2.2 ppm; S02 <0.01 ppm) for 1 day or 1, 4, or 8 weeks (7/h/day, 5 days/week, endpoints
12-h post DE exposure). Results from the 1- and 4-week exposures are described in Section 6.3.6.2. It
should be noted that control mice were left in a clean room as opposed to undergoing chamber exposure
to filtered air. The significant increases in AHR and airway sensitivity observed following shorter
exposure periods did not persist at 8 weeks. BALF cytokines were altered by DE exposure with only
RANTES significantly elevated after 8 weeks. DE had no effect on OVA challenge-induced peribronchial
inflammatory or mucin positive cells. These results suggest that adaptive processes may have occurred
during prolonged exposure to DE.
Woodsmoke
In a study by Tesfaigzi et al. (2005), Brown Norway rats were sensitized and challenged with
ovalbumin. Rats were exposed for 70 days to filtered air or to 1000 (ig/m3 HWS. Particles were
characterized by a MMAD of 0.36 |_im. Concentrations of gases were reported to be 13.0 ppm CO and 3.1
ppm total vapor hydrocarbon with negligible NOx. Respiratory function was measured in anesthetized
December 2008
7-46
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
animals by whole-body plethysmography and demonstrated a significant increase in functional residual
capacity as well as a significant increase in dynamic lung compliance in HWS-exposed animals compared
to controls. No change in total pulmonary resistance or airway responsiveness to methacholine was
observed. BALF inflammatory cells were not increased, although histological analysis demonstrated focal
inflammation including granulomatous lesion and eosinophilic infiltrations in HWS-exposed rats.
Alterations of several cytokines in BALF and plasma were noted. Changes in airway epithelial mucus
cells and intraepithelial stored mucosubstances were modest and did not achieve statistical significance.
Results of this study demonstrate that subchronic exposure to HWS had minimal effects on pulmonary
responses in a rat model of allergen sensitization and challenge.
7.3.7. Host Defense
7.3.7.1. Toxicological Studies
Diesel Exhaust
DE may affect systemic immunity. The proliferative response of A/J mouse spleen cells following
stimulation with T cell mitogens was suppressed by 6 months of daily exposure to DE at concentrations at
or above 300 (ig/m3 PM (Burchiel et al., 2004). B cell proliferation was increased at 300 (ig/m3 but
unaffected at higher concentrations (up to 1000 (.ig/ni3). Concentrations of gases were not reported. The
immunosuppressive effects of DE were not due to PAHs or benzo(a)pyrene (BaP)-quinones (BPQs) since
there were little, if any, of these compounds present in the chamber atmosphere. It should be noted that
sentinel animals were negative for mouse parvovirus at the start of the study, but seroconverted by the end
of the study, indicating possible infection. Parvovirus can interfere with the modulation of lymphocyte
mitogenic responses (Baker, 1998).
Woodsmoke
One study demonstrated immunosuppressive effects of HWS (Burchiel et al., 2005). Exposure to
HWS increased proliferation of T cells from A/J mice exposed daily to 100 (ig/m3 PM for 6 months, but
produced a concentration-dependent suppression of proliferation at PM concentrations >300 (ig/m3. No
effects on B cell proliferation were observed. Concentrations of NO and N02 were not detectable or <40
ppb for all exposure levels. CO was reported to be 2, 4, and 13 ppm for the 100, 300 and 1000 (ig/m3 PM
concentrations, respectively. Exposure atmospheres contained significant levels of naphthalene and
methylated napthalenes, fluorene, phenanthrene, and anthracene, as well as low concentrations of several
metals (K, Ca, and Fe) (Burchiel et al., 2005). It should be noted that serologic analysis of study sentinel
December 2008
7-47
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
animals indicated infection with parvovirus, which can interfere with the modulation of lymphocyte
mitogenic responses (Baker, 1998).
7.3.8. Summary and Causal Determination
7.3.8.1. PM10
Recent long-term, prospective cohort studies support the positive association between respiratory
symptoms and ambient PMi0 concentrations reported in the 2004 PM AQCD. This includes evidence for a
reduction of symptoms corresponding to decreasing PMi0 levels in a "natural experiment" in a cohort of
Swiss school children. Epidemiologic studies provide evidence of a consistent and coherent relationship
between PMi0 levels and decrements in lung function and lung function growth (Figure 7.4), confirming
the relationship reported in the 2004 PM AQCD. Overall, the evidence is sufficient to conclude that
the relationship between long-term PM10 exposure and respiratory morbidity is likely to be causal.
The epidemiologic studies reviewed in the 2004 PM AQCD suggested positive relationships
between long-term PMi0 exposure and increased incidence of respiratory symptoms and disease. The
expanded body of evidence available now includes prospective cohort studies conducted by different
researchers in different locations. In the U.S., studies include those from the CHS cohort (e.g., Islam et
al., 2007; McConnell et al., 2003), the national cystic fibrosis study (Goss et al., 2004) and a study in San
Francisco by Kim et al. (2004). These, along with studies conducted in Europe, build upon the evidence
available in the 2004 PM AQCD. Bayer-Oglesby et al. (2005) provide evidence, in a cohort consisting of
school children, for respiratory symptoms reduction where PM levels are decreasing. A major challenge to
interpreting the results of these studies is that the PM measures and concentrations of other air pollutants
are often correlated; however, the consistency of findings across different locations supports an
independent effect of PMi0. Overall there is a growing database relating respiratory symptoms and disease
incidence to long-term PMi0 exposure in U.S. communities that is supported by similar results from other
studies outside the U.S.
The 2004 PM AQCD also suggested positive relationships for pulmonary function declines with
long-term exposure to PMi0. Recent studies have been conducted in the U.S. with the national cystic
fibrosis cohort by Goss et al. (2004) and the CHS by Gauderman et al. (2004). These build on the
database for the U.S. of the previously reviewed studies in other U.S. locations (Raizenne et al., 1996)
and on earlier CHS results (e.g., Avol et al., 2001; Gauderman et al., 2000; 2002), showing a pattern of
reductions in lung function with increased long-term PMi0 exposures. These are supported by other
longitudinal cohort studies conducted in other countries (Nordling et al., 2008; Oftedal et al., 2008; Rojas-
December 2008
7-48
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Martinez et al., 2007). In addition, Downs et al. (2007) report an association between a decrease in PM
and pulmonary function declines in a cohort of adults. The studies provide a consistent and coherent
relationship between PM exposure and change in lung function. However, the strong correlation between
PM and other pollutants again complicates the identification of PM as an independent causal factor.
7.3.8.2. PM2.5
The 2004 PM AQCD suggested a positive relationship between PM2 5 and lung function
decrements reported in the CHS and increased bronchitis. Recent epidemiologic studies support these
findings and provide additional evidence of associations between PM2 5 and lung function decrements and
increased respiratory symptoms/asthma medication use. In the 2004 PM AQCD, no long-term exposure
toxicological studies were available. Recent subchronic and chronic toxicological studies provide some
evidence of altered pulmonary function, mild inflammation, oxidative responses, histopathological
changes including mucus cell hyperplasia and immune suppression in response to CAPs, DE, roadway air
and woodsmoke. Allergic animals demonstrated AHR in response to DE. In some cases, adaptation to
prolonged exposures was observed. In addition, pre- and postnatal exposure to ambient levels of urban
particles was found to affect mouse lung development. Impaired lung development is one mechanism by
which PM exposure may decrease lung function growth in children. Collectively, the evidence is
sufficient to conclude that the relationship between long-term PM2.5 exposure and respiratory
morbidity is likely to be causal.
Respiratory symptoms/disease
The epidemiologic studies reviewed in the 2004 PM AQCD suggested relationships between PM2 5
(PM21) and bronchitis in the 24-city cohort of Dockery et al. (1996). In the U.S., new PM2 5 studies show
associations with respiratory symptoms and with asthma (Goss et al., 2004; Islam et al., 2007; Kim et al.,
2004; McConnell et al., 2003). These are supported by studies outside the U.S. (Annesi-Maesano et al.,
2007; Brauer et al., 2007). Brauer et al. (2007) confirm the results of Islam et al. (2007) for the
relationship between asthma and PM. Because they represent a different research group, in a different
cohort and a different design and analysis, their results add to the strength of the observations. The within-
community results of McConnell et al. (2003) for bronchitic symptoms were larger than the between-
community results. There are less likely confounding effects because of the within community aspect.
This provides direct evidence that reduction in air pollutants could result in improved respiratory health in
children.
December 2008
7-49
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Pulmonary Function
The epidemiologic studies reviewed in the 2004 PM AQCD suggested relationships between long-
term exposure to PM2 5 and pulmonary function decrements in the CHS (Gauderman et al., 2000; 2002)
which added to the database of the earlier 22-city study of PM21 (Raizenne et al., 1996). In the U.S., new
PM2 5 studies show associations with reduced pulmonary function (Gauderman et al., 2004; Goss et al.,
2004; Islam et al., 2007). These are supported by studies outside the U.S. (Dales et al., 2008; Oftedal et
al., 2008). A subset of the PM2 5 measurements in the CHS also show associations for EC, OC, and acid
vapor as presented by Gauderman et al. (2004) and Islam et al. (2007).
Recent toxicological studies demonstrate alterations in pulmonary function; AHR was increased
following 4, but not 8, weeks of exposure to low dose DE in one of two non-allergic mouse strains (Li et
al., 2007). Adaptation may have occurred during the prolonged exposure. Total pulmonary resistance was
increased in rats exposed for 4 and 12 weeks to woodsmoke (Tesfaigzi et al., 2002). In addition, an
important new study demonstrated that pre- and postnatal exposure to ambient levels of urban particles
affected mouse lung development, as measured by anatomical and functional indices (Mauad et al., 2008).
These results provide biological plausibility for the epidemiologic findings of reduced lung function
growth in children.
Pulmonary Inflammation
One epidemiologic study found an increase in eNO among schoolchildren (Dales et al., 2008).
Recent toxicological studies demonstrate variable inflammatory responses with PM exposure.
Toxicological studies involving subchronic exposure to CAPs found little or no pulmonary inflammation
(Araujo et al., 2008; Lippmann et al., 2005a). Pulmonary inflammation was observed in 3 out of 5 studies
involving subchronic exposure to DE (Hiramatsu et al., 2003; Ishihara and Kagawa, 2003; Li et al., 2007).
One of the three studies with positive results found that gaseous components of DE had minimal effects
(Ishihara and Kagawa, 2003). Subchronic woodsmoke exposure resulted in pulmonary inflammation in 2
(Seagrave et al., 2005b; Tesfaigzi et al., 2002) out of 3 studies.
Oxidative Responses
One new toxicological study demonstrated increases in lung HO-1 protein and mRNA levels
following 4 weeks of DE exposure in two mouse strains (Li et al., 2007). This effect persisted over 8
weeks of DE exposure in one strain but not the other.
December 2008
7-50
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Pulmonary Injury
Recent toxicological studies demonstrate mild injury as assessed by BALF markers and
histopathological responses. Injury was observed in rats exposed for 12 months to DE (Ishihara and
Kagawa, 2003). This effect was attenuated when the exhaust was filtered to remove PM.
Histopathological changes were observed in some long-term studies. Nasal and airway mucous cell
hyperplasia was seen in mice following long-term exposure to urban air from a heavily-trafficked area.
These changes were accompanied by an increase in total and acidic mucus production which can lead to a
loss of mucus-mediated protective functions (Pires-Neto et al., 2006). Long-term exposure of rats to
roadside air resulted in mast cell infiltration and hypertrophy of alveolar walls (Kato and Kagawa, 2003).
Airway mucous cell hyperplasia was observed in rats following long-term woodsmoke exposure
(Tesfaigzi et al., 2002). Both neutral and acidic mucus producing cells were increased in number. In
addition, an important new study demonstrated that pre- and postnatal exposure to ambient levels of urban
particles affected mouse lung development, as measured by changes in pulmonary surface to volume
ratios and by changes in lung function (Mauad et al., 2008). A study reviewed in Section 6.3.4.2 also
suggested that the developing lung may be susceptible to PM since acute exposure to ultrafine-soot PM
decreased cell proliferation in the proximal alveolar region of neonatal rats (Pinkerton et al., 2004).
Impaired lung development is a viable mechanism by which PM may reduce lung function growth in
children.
Allergic Responses
Recent toxicological studies suggest that adaptive processes may have occurred during prolonged
exposures. Allergic mice exposed for 8 weeks to DE did not exhibit the AHR response observed at 1 and
4 weeks of exposure (Matsumoto et al., 2006). A long-term exposure study of woodsmoke exposure in
allergic rats demonstrated no increased AHR although focal inflammation and eosinophilic infiltration
was observed (Tesfaigzi et al., 2005).
Host Defense
Recent toxicological studies demonstrate immune suppression, indicated by decreased proliferative
responses of T-cells in spleen, in mice exposed chronically to DE and woodsmoke (Burchiel et al., 2004;
Burchiel et al., 2005). In another study, exposure to woodsmoke did not alter bacterial clearance in
response to bacterial infection (Reed et al., 2006).
December 2008
7-51
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
7.3.8.3. Ultrafine PM
The evidence is inadequate to determine if a causal relationship exists between relevant UFP
exposures and long-term respiratory morbidity. The 2004 PM AQCD did not report long-term
exposure studies for ulrafine PM. More recent studies involving subchronic exposure to ultrafine CAPs
found no major pulmonary inflammation (Araujo et al., 2008).
7.4. Reproductive, Developmental, Prenatal and Neonatal
Outcomes
7.4.1. Epidemiologic Studies
This section evaluates and summarizes the scientific evidence on PM and developmental and
pregnancy outcomes and infant mortality. Infants and fetal development processes may be particularly
vulnerable to PM exposure, and although the physical mechanisms are not fully understood, several
hypotheses have been proposed involving direct effects on fetal health, altered placenta function, or
indirect effects on the mother's health (Bracken et al., 2003; Clifton et al., 2001; Glinianaia et al., 2004;
Maisonet et al., 2004; Schatz et al., 1990; Sram et al., 2005). Study of these outcomes can be difficult
given the need for detailed data and potential residential movement of mothers during pregnancy. Two
recent articles have reviewed methodological issues relating to the study of outdoor air pollution and
adverse birth outcomes (Ritz and Wilhelm, 2008; Slama et al., 2008). Some of the key challenges to
interpretation of these study results include the difficulty in assessing exposure as most studies use
existing monitoring networks to estimate individual exposure to ambient PM; the inability to control for
potential confounders such as other risk factors that affect birth outcomes (e.g., smoking); evaluating the
exposure window (e.g., trimester) of importance; and the need to study the physiological mechanism of
these effects (Ritz and Wilhelm, 2008; Slama et al., 2008). Another uncertainty is whether PM effects
differ by the child's sex. A review of pre-term birth and low birth weight studies found limited indication
that effects may differ by gender, however sample size was limited (Ghosh et al., 2007).
Previous summaries of the association between PM concentrations and pregnancy outcomes and
infant mortality were presented in periodic PM AQCDs. The 1996 PM AQCD concluded that while few
studies had been conducted on the link between PM and infant mortality, the research "suggested an
association," particularly for post-neonates (U.S. EPA, 1996). At the writing of the 2004 PM AQCD,
several studies conducted since the 1996 version provided additional evidence of PM's effect on fetal and
December 2008
7-52
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
early postnatal development and mortality (U.S. EPA, 2004). At that time, it was concluded that while
some studies indicated a relationship between PM and pregnancy outcomes, others did not. Studies
identifying associations found that exposure to PM10 early during pregnancy (first month of pregnancy) or
late in the pregnancy (six weeks prior to birth) were linked with higher risk of preterm birth, including
models adjusted for other pollutants, and that PM2.5 during the first month of pregnancy was associated
with interuterine growth restriction. However, other work did not identify relationships between PMi0
exposure and low birth weight. The state of the science at that time, as indicated in the 2004 PM AQCD,
was that the research provided mixed results based on studies from multiple countries, and that additional
research was required to better understand the impact of PM on pregnancy outcomes and infant mortality.
The relationship between PM and these health responses, with particular emphasis on findings since the
previous PM document, are summarized here. Over all, epidemiologic studies consistently report
associations between PMi0 and PM2 5 exposure and low birth weight and infant mortality, especially
during the post-neonatal period. Animal evidence supports these associations with PM2 5, but provides
little mechanistic information or biologic plausibility. Information on the ambient concentrations of PMi0
and PM2 5 in these study sites can be found in Table 7-5.
Low Birth Weight
A large number of studies have investigated exposure to ambient PM and low birth weight at term.
After briefly reviewing three older studies that measured total suspended particulates (TSP), this review
will concentrate on studies that measured PM distinguished by size fraction.
In 1997, Wang et al. reported increased risks for low birth weight related to TSP exposure in
Beijing, China (Wang et al., 1997). Women in four residential districts in Beijing were studied. Exposure
was analyzed in quintiles and women in the fourth (437-497 |ig/m3) and fifth (498-618 |ig/m3) quintiles
were at increased risk for low birth weight (OR = 1.15 [95%CI: 1.00-1.32], OR = 1.24 [95%CI: 1.08-
1.42) compared to the lowest quintile. These are very high exposures and substantially in excess of
current EPA standards, however, the comparison group was exposed to 211-280 |ig/m3. Actual risks of
exposure may have been much higher, if compared to an unexposed group.
Lower levels of TSP were reported by Bobak in the Czech Republic (25th percentile = 54.8 |ig/m3.
50th percentile = 71.5 (ig/m3, 75th percentile = 86.9) (Bobak and Leon, 1999b). A 50 |ig/m3 increase in
TSP was not associated with increased risk for low birth weight when the model was adjusted for
gestational age. In Seoul, Korea slightly higher levels of TSP (25th percentile = 76.7 (ig/m3, 50th
percentile = 82.3 |ig/m3. 75th percentile = 91.0 (.ig/nr3) were associated with increased risks of low birth
weight for exposure in the first trimester (OR = 1.04 [95% CI: 1.00-1.08]), but not in the third trimester
(OR = 0.95 [95% CI: 0.90-0.99]) (Haetal., 2001).
December 2008
7-53
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
More recent studies of the effect of PM on birth weight in the U.S. may be more relevant to EPA
regulation. Since 2000 one national study, as well as two studies in the northeast U.S., and four in
California have been conducted. Parker and Woodruff (2008) linked U.S. birth records for singletons
delivered at 40 weeks gestation in 2001-2003 during the months of March, June, September and
December to quarterly estimates of PM exposure by county of residence and month of birth. They found
an association between PM10-2.5 and birthweight (-13 g [95% CI: -18.3 to -7.6]) per 10 |ig/m3 increase),
but no such association for PM2 5.
Maisonet et al. (2001) analyzed 89,557 births (1994-96) in six northeastern cities (Boston and
Springfield MA, Hartford CT, Philadelphia and Pittsburgh PA, Washington DC). Each city had three PMi0
monitors measuring every sixth day. Results from multiple monitors were averaged in each city. Exposure
was determined for each trimester of pregnancy and categorized by quartiles, (<25, 25-30, 31-35, 36-
43 |ig/m3) and 95th percentile (>43|ig/m3). There was no increased risk for low birth weight at term
associated with PMi0 exposure during any trimester of pregnancy. When birth weight was considered as a
continuous outcome, exposure to PMi0 was not associated with a reduction in mean birth weight.
Table 7-5. Characterization of ambient PM concentrations from studies of reproductive,
developmental, prenatal and neonatal outcomes and long-term exposure.
Reference
Location
Mean Annual
Concentration (|jg/m3>
Upper Percentile
Concentrations (|jg/m3'
PMw
Bell etal. (2007b)
CT & MA
22.3
Brauer et al. (2008)
Vancouver, Canada
12.7
Max: 35.4
Chen et al. (2002)
Washoe County, NV
31.53
75th: 39.35; Max: 157.32
Gilboa et al. (2005)
TX
23.8=
75th: 29
Ha et al. (2003)
Seoul, South Korea
69.2
75th: 87.7; Max: 245.4
Hansen et al. (2006)
Brisbane, Australia
19.6
Max: 171.7
Hansen et al. (2007a)
Brisbane, Australia
19.6
75th: 22.7; Max: 171.7
Jalaludin et al. (2007)
Sydney, Australia
16.3
Kim et al. (2007b)
Seoul, Korea
88.7-89.7
Lee et al. (2003a)
Seoul, Korea
71.1
75th: 89.3; Max: 236.9
Leem et al. (2006)
Incheon, Korea
53.8=
75th: 64.6; Max: 106.39
Lipfert et al. (2000)
U.S.
33.1
Max: 59
Maisonet et al. (2001)
NE U.S.
31.0=
75th: 36.1; Max: 46.5
Mannes et al. (2005)
Sydney, Australia
16.8
75th: 19.9; Max: 104.0
Pereira et al. (1998)
Sao Paulo, Brazil
65.04
Max: 192.8
Ritz et al. (2000)
CA
49.3
Max: 178.8
Ritz et al. (2006)
CA
46.3
Max: 83.5
Rogers and Dunlop (2006)
GA
3.2-7.8=
75th: 15.07
Romieu et al. (2004)
Ciudad Juarez, Mexico
33.0-45.9
December 2008
7-54
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Salam et al. (2005) CA 45.4-46.6
Tsai et al. (2006b) Kaohsiung, Taiwan 81.5 75th: 111.5; Max: 232.0
Wilhelm and Ritz (2005) Los Angeles, CA 39.1-42.2 Max: 74.6-103.7
Woodruff et al. (2008) U.S. 28.6-29.8= 75th: 33.8-36.5
Yang et al. (2006) Taipei, Taiwan 53.2 75th: 64.9; Max: 234.9
PMis
Basuetal. (2004) CA 14.5-18.2 Max: 26.3-34.1
Bell etal. (2007b) CT&MA 22.3
Brauer et al. (2008) Vancouver, Canada 5.3 Max: 37.0
Huynh et al. (2006) CA 17.5-18.8
Jalaludin etal. (2007) Sydney, Australia 9.0
Liu et al. (2007b) Multicity, Canada 12.2 75th: 15
Loomis et al. (1999) Mexico City 27.4 Max: 85
Mannes et al. (2005) Sydney, Australia 9.4 75th: 11.2; Max: 82.1
Parker etal. (2005) CA 15.4
Ritz et al. (2007) Los Angeles, CA 18.6-21.4
Sagiv et al. (2005) PA 25.3-27.1 Max: 68.9-156.3
Wilhelm and Ritz (2005) Los Angeles, CA 21.0 Max: 38.9-48.5
Woodruff etal. (2006) CA 19.2= 75th: 22.7
Woodruff et al. (2008) U.S. 14.5-14.9 75th: 18.5-18.7
"Median concentration
In contrast, Bell et al. (2007b) reported positive associations for both PMi0 and PM2 5 with birth
weight in a study of births (n = 358,504) in Connecticut and Massachusetts (1999-2002). Birth data
indicated county, not street address or ZIP code, so women were assigned exposure based on county
residence at delivery. The difference in birth weight per 10 (.ig/nr1 associated with PM2 5 was -66.8 (95%
CI: 77.7 to -55.9) gm. For PMi0 it was -11.1 (95% CI: -15.0 to -7.2) gm. The increased risk for low birth
weight was OR= 1.054 (95% CI: 1.022-1.087) for PM2 5 and OR= 1.027 (95% CI: 0.991-1.064) for
PM10, based on average exposure during pregnancy. Reductions in birth weight were also associated with
third trimester exposure to PMi0 and second and third trimester exposure to PM2 5. Comparing this study
to Maisonet et al. (2001), a larger sample size was able to detect a small increase in risk. In addition, birth
weight was reduced more by exposure to PM2 5 than by exposure to PMi0. Measured PM2 5 concentrations
were not available in the earlier study.
The Children's Health Study is a population based cohort of children living in 12 southern
California communities, selected on the basis of differing levels of air pollution (Salam et al., 2005). The
children in grades 4, 7 and 10 were recruited through schools. A subset of this cohort (n = 6,259) were
born in California from 1975-1987. Of these, birth certificates were located for 4,842, including 3,901
infants born at term and 72 cases of low birth weight at term. Using the mother's ZIP code at the time of
birth, exposure was determined by inverse distance weighting of up to 3 PMi0 monitors within 50 km of
the ZIP code centroid. If there was a PMi0 monitor within 5 km of the ZIP code centroid (40% of data),
December 2008
7-55
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
exposure from that monitor was used. Exposure was calculated for the entire pregnancy, and for each
trimester of pregnancy. A 10 |ig/m3 increase in PM10 during the third trimester reduced mean birth weight
-10.9 g (95% CI: -21.1 to -0.6) in single pollutant models, but became non-significant in multipollutant
models controlling for the effects of 03. Increased risks of low birth weight (<2500 gm) were not
statistically significant (OR =1.3 [95% CI: 0.9-1.9]). A strength of this study was the cohort data
available included information on SES and smoking during pregnancy. A limitation is the assignment of
exposure based on monitoring stations up to 50 km distant; this may have introduced significant
misclassification obscuring some associations.
Parker et al. examined births in California within 5 miles of a monitoring station (n = 18,247)
(Parker et al., 2005). Only infants born at 40 weeks gestation were included. Thus all infants were the
same gestational age, and had been exposed in the same year. Exposure to PM2.5 in quartiles (<11.9,
11.9-13.9, 14.0-18.4, >18.4) was associated with decrements in birth weight. Infants exposed to
>13.9 |ig/m3 experienced reductions in birth weight (third quartile -13.7 g (95% CI: -34.2 to 6.9), fourth
quartile -36.1 g (95% CI: -55.8 to -16.5). These are larger reductions than have been seen in some other
studies. However, this study reduced misclassification by including only women living within 5 miles of a
monitoring station, and only included births at 40 weeks gestation. Reducing misclassification should lead
to a stronger association, if the association is causal.
The effects of spatial variation in exposure were also investigated by Wilhelm and Ritz (2005).
Their study included all women living in ZIP codes where 60% of the ZIP code was within two miles of a
monitoring station in the Southern California Basin, and women with known addresses in Los Angeles
County within 4 miles of a monitoring station. Exposure to average PM10 in the third trimester was
analyzed for increased risk of low birth weight at term (> 37 weeks gestation). Analysis at the ZIP code
level did not detect increased risk (per 10 |ig/m3 PMi0, OR= 1.03 [95% CI: 0.97-1.09]). However the
analysis based on geocoded addresses indicated that increasing exposure to PMi0 was associated with
increased risk of low birth weight for women living within 1 mile of the station where PMi0 was
measured. For these women (n = 247 cases, 10,981 non-cases), each 10 |ig/m3 increase in PMi0 was
associated with a 22% increase in risk of term low birth weight (OR = 1.22 [95% CI: 1.05-1.41]). In the
categorical analysis, exposure to PMi0 >44.4 |ig/m3 was associated with a 48% increase in risk
(OR = 1.48 [95% CI: 1.00-2.19]). Increased risk of low birth weight also was associated with exposure to
CO in single pollutant models. However, when multipollutant models were considered, the effects of CO
were attenuated but the effects of PMi0 increased. Controlling for CO, N02, and 03, each 10 |ig/m3
increase in exposure to PMi0 increased risk of low birth weight 36% (OR = 1.36 [95% CI: 1.12-1.65]).
Spatial variation in PM2 5 exposure was investigated by Basu et al. (2004). They included only
mothers who lived within 5 miles of a PM2 5 monitor and within a California county with at least one
December 2008
7-56
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
monitor. To minimize potential confounding, they included only white (n = 8597) or Hispanic (n = 8114)
women, who were married, between 20-30 years of age, completed at least high school and were having
their first child. Consistently, PM2 5 exposure measured by the county monitor was more strongly
associated with reductions in birth weight than exposure measured by the neighborhood monitor. The
results were replicated in both the white and the Hispanic samples. Reductions in birth weight ranged
from 15.2 g to 43.5 g per 10 |ig/m3 increase in PM25.
In the remaining U.S. study, Chen et al. (2002) analyzed 33,859 birth certificates of residents of
Washoe County in northern Nevada (1991-1999). There were four sites monitoring PMi0 during the study
period, it appears (not stated) that exposure was averaged over the county. A 10 |ig/m3 increase in
exposure to PMi0 during the third trimester of pregnancy was associated with an 11 gm reduction in birth
weight (95% CI: -2.3 to -19.8). Effects on risk of low birth weight were not significant. For exposure in
the third trimester of 19.77 to 44.74 (ig/m3 compared to <19.74 |ig/m3 the odds ratio for low birth weight
was 1.05 (95% CI: 0.81-1.36). Comparing exposure >44.74 to the same reference category, the odds ratio
was 1.10 (95% CI: 0.71-1.71). Misclassification of exposure may have occurred when exposure was
averaged over a large geographic area (16,968 km2).
Recent international studies investigating effects of particles on low birth weight include one in
Munich (Slama et al., 2007), two in Canada (Brauer et al., 2008; Dugandzic et al., 2006), two in Australia
(Hansen et al., 2007a; Mannes et al., 2005), two in Taiwan (Lin et al., 2004b; Yang et al., 2003a), one in
Korea (Ha et al., 2003) and two in Sao Paulo, Brazil (Gouveia et al., 2004; Medeiros and Gouveia, 2005).
The majority of these studies found that PM concentrations were associated with low birth weight, though
two studies (Hansen et al., 2007a; Lin et al., 2004b) found no associations. The effect estimates were
similar in magnitude to those reported in the U.S. studies.
Issues in Interpreting Results of Low Birth Weight Studies
Studies included subjects at distances from monitoring stations varying from as close as 1 mile or 2
km, to as far as 50 km or the size of the county. However, studies that only included subjects living within
a short distance (1 mile, 2 km) of the monitoring station (thus likely reducing exposure measurement
error) were more likely to find that exposure was associated with increased risk of low birth weight.
However, Basu et al. (2004) reported a stronger association between PM2 5 exposure and birth weight
when exposure was estimated based on the county monitor, rather than the monitor within 5 miles of the
residence. They suggest that county level exposure may be more representative of where women spend
their time, including not only home, but also other time spent away from home. Other pollutants also
appeared to influence the risk associated with particle exposure. In one study, exposure to PMi0 in a single
pollutant model reduced birth weight by 11 grams, but became non-significant in multipollutant models
December 2008
7-57
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
with 03 (Salam et al., 2005). In another study the risk associated with PM10 exposure increased from 22%
to 36% when other pollutants were included in the model (Wilhelm and Ritz, 2005). All but one study in
the U.S. found some association between particle exposure and reduced birth weight (Maisonet et al.,
2001). The results of international studies were mixed. This might be related to the chemical composition
of particles in the U.S., or to differences in the pollutant mixture. Studies with negative results must be
interpreted with caution when the comparison groups have significant exposure. This was certainly the
situation in studies in Taiwan and Korea (Lee et al., 2003a; Lin et al., 2004b; Yang et al., 2003a).
Differences in geographical locations, study samples and linkage decisions may contribute to the diverse
findings in the literature on the association between PM and birthweight, even within the U.S. (Parker and
Woodruff, 2008).
Preterm Birth
A potential association of exposure to airborne particles and preterm birth has been investigated in
numerous epidemiologic studies, including some conducted in the U.S. and others in foreign countries.
Three U.S. studies have been carried out by the same group of investigators in California.
A natural experiment occurred when an open-hearth steel mill in Utah Valley was closed from
August 1986 through September 1987. Parker et al. (2008) compared birth outcomes for Utah mothers
within and outside of the Utah Valley, before, during, and after the mill closure. They report that mothers
who were pregnant around the time of the closure of the mill were less likely to deliver prematurely than
mothers who were pregnant before or after. The strongest effect estimates were observed for exposure
during the second trimester (14% decrease in risk of preterm birth during mill closure). Preterm birth
outside of the Utah Valley did not change during the time of the mill closure.
In 2000, Ritz et al. (2000) published the first study investigating the association of preterm birth
with PM in the U.S. The study population was women living in the southern California Basin. There were
8 monitoring stations measuring PMi0 every sixth day during the study period. Birth certificates (1989-
1993) were analyzed for women living in zip codes within 2 miles of a monitoring station. Women with
multiple gestations, chronic disease prior to pregnancy and women who delivered by cesarean section
were excluded resulting in a study population of 48,904 women. The risk of preterm birth increased by
4% (RR = 1.04 [95% CI: 1.02-1.6]) per 10 (ig/m3 increase in PMi0 averaged in the six weeks before birth.
Exposure to PMi0 in the first month of pregnancy, resulted in a 3% increase in risk (RR = 1.03 [95% CI:
1.01-1.05]). These results were robust in multipollutant models.
Wilhelm and Ritz (2005) reinvestigated this association among women in the same area in 2005,
when air pollution had declined. Birth certificate data from 1994-2000 was analyzed for women living in
ZIP codes within 2 miles of a monitoring station, or with addresses within 5 miles of the monitoring
December 2008
7-58
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
station. No significant effects of exposure to PM10 were reported. Exposure to PM2 5 six weeks before
birth resulted in an increase in preterm birth (RR =1.19 [95% CI: 1.02-1.40]) for the highest quartile of
exposure (PM25 >24.3 |ig/m3). Using a continuous measure of PM25, there was a 10% increase in risk for
each 10 ^ig/m3 increase in PM2 5 (RR= 1.10 [95% CI: 1.00-1.21]).
There have been two major criticisms of air pollution studies using birth certificate data. First, that
birth certificates only indicate the address at birth and the exposure of women who moved during
pregnancy may be misclassified; second, that information about some important confounders may not be
available (e.g. smoking). To obtain more precise information about these variables, Ritz et al. (2007)
conducted a case control study nested within a cohort of birth certificates (Jan 2003-Dec 2003) in Los
Angeles county. Births to women residing in ZIP codes (n = 24) close to monitoring stations or major
population centers or roadways (n = 87) were eligible (n = 58,316 births). All cases of low birth weight or
preterm birth and an equal number of randomly sampled controls in the 24 zip codes close to monitors
were selected. In the other 87 ZIP codes, 30% of cases and an equal number of controls were randomly
sampled. Of 6,374 women selected for the case control study, 2,543 (40%) were interviewed. The
association of preterm birth with exposure to PM2 5 differed between women responding to the survey and
women who did not respond. Among responders, exposure to each 10 (ig/m3 increase in PM2 5
concentration in the first trimester increased risk to preterm birth by 23% (RR = 1.23 [95% CI:
1.02-1.48]). There was no increase in risk among non-responders (RR= 0.95 [95% CI: 0.82-1.10]), or in
the entire birth cohort (RR= 1.00 [95% CI: 0.94-1.07]).
An additional case control study of preterm birth and PM2 5 exposure (Huynh et al., 2006) used
California birth certificate data. Singleton preterm infants (24-36 weeks gestation) born in California
(1999-2000) whose mothers lived within 5 miles of a PM2 5 monitor were eligible. Each of these 10,673
preterm infants were matched to three term (39-44 weeks gestation) controls (having a last menstrual
period within 2 weeks of the case infant), resulting in a study population of 42,692. Controlling for
maternal race/ethnicity, education, marital status, parity and CO exposure, exposure to PM2 5 >17.7 (ig/m3
increased the risk of preterm birth by 14%(OR= 1.14 [95% CI: 1.07-1.23]). Averaging PM25 exposure
over the first month of pregnancy, the last 2 weeks before birth, or the entire pregnancy did not
substantially change the risk estimate.
Two additional studies of preterm birth and exposure to particulate air pollution have been
conducted in the U.S. Each has used a unique methodology. Sagiv et al. (2005) used time series to analyze
births in four Pennsylvania counties between January 1997 and December 2001. In this analysis, exposure
is compared to the rate of preterm births each day. Both acute exposure (on the day of birth) and longer
term exposure (average exposure for the preceding six weeks) were considered in the analysis. An
advantage of this analysis is that days, rather than individuals are compared, so confounding by individual
December 2008
7-59
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
risk factors is minimized. For exposure averaged over the six weeks prior to birth, there was a non-
significant increase in risk (RR = 1.07 [95% CI: 0.98-1.18]); for acute exposure with a 2 day lag
(RR= 1.10 [95% CI: 1.00-1.21]) and 5 day lag (RR= 1.07 95% CI: 0.98-1.18]) results were marginal.
Rogers and Dunlop (2006) examined exposure to particles and risk of delivery of an infant
weighing less than 1500 grams (all of which were preterm) from 24 counties in Georgia. The study
included 69 preterm small for gestational age (SGA) infants, 59 preterm appropriate for gestational age
(AGA) infants and 197 term AGA controls. Exposure was estimated using an environmental transport
model that considered PMi0 emissions from 32 geographically located industrial point sources,
meteorological factors, and geographic location of the birth home. Exposure was categorized by quartiles.
Comparing women who delivered a preterm AGA infant to those who delivered a term AGA infant,
exposure to PMi0>15.07 (ig/m3 tripled the risk (OR = 3.68 [95% CI: 1.44-9.44]).
Brauer et al. (2008) evaluated the impacts of PM2 5 on preterm birth using spatiotemporal exposure
metrics in Vancouver, Canada. The authors found similar results when they used a land-use regression
model or inverse distance weighting as the exposure metric. For preterm births <37 weeks, they reported
an OR of 1.06 (95% CI: 1.01-1.11), and for preterm births <35 weeks the OR increased to 1.12 (95% CI:
1.02-1.24). There were no consistent trends for early or late gestational period to be more strongly
associated with preterm births.
In Incheon, Korea, Leem et al. (2006) estimated PM10 exposure spatially as well as temporally.
Exposure was based on 26 monitors and kriging was used to determine exposure for 120 dongs
(administrative districts, mean area 7.82 km2, median area 1.42 km3). The sample included 52,113 births,
from 2001-2002. PM10 was very weakly correlated with other pollutants. Exposure was compared in
quartiles for the first and third trimester of pregnancy. In the first trimester, relative risks for the second,
third and fourth quartiles were RR= 1.14 (95% CI: 0.97-1.34), RR= 1.07 (95% CI: 0.94-1.37), and
RR = 1.24 (95% CI: 1.09-1.41), respectively. Exposure to PMi0 in quartile one (reference group) was 26.9
- 45.9 |ig/m3: fourth quartile exposure equaled 64.6-106.4 |ig/nr\ The p-value for trend was 0.02.
Exposure in the third trimester was not related to preterm birth, however no information was provided to
determine how exposure in the third trimester was adjusted for women who delivered preterm.
Two studies investigating risks of preterm birth related to particle exposure have been reported
from Australia. In Brisbane, Hansen et al. (2006) studied 28,200 births (2000-2003) in an area of low
PMio concentrations. Exposure to an interquartile range increase in PMi0 exposure in the first trimester
resulted in a 15% increased risk of preterm birth (OR =1.15 [95% CI: 1.06-1.25]). This result was
strongly influenced by the effect of PMi0 exposure in the first month of pregnancy (OR =1.19 [95% CI:
1.13-1.26]). PMio was correlated with ozone r = (0.77) in this study and ozone also increased risk in the
first trimester. No effects were associated with exposure to PMi0 in the third trimester.
December 2008
7-60
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
In Sydney, associations between exposure to particles and preterm birth varied by season. Jalaludin
et al. (2007) obtained information on all births in metropolitan Sydney (1998-2000). Exposure to PM2 5 in
the three months preceding birth was associated with an increased risk of preterm birth (OR =1.11 [95%
CI: 1.04-1.19]). Additional effects were dependent on season of conception. Both PMi0 (OR= 1.3 95%
CI: 1.2-1.5]) and PM2 5 (OR= 1.4 [95% CI: 1.3-1.6]) were associated with increased risk for conceptions
in the winter. Conceptions in summer were associated with reductions in risk (PMi0 OR = 0.91 [95% CI:
0.88-0.93]), (PM2 5 OR = 0.87 [95% CI: 0.84-0.92]). Due to both positive and negative findings, the
authors recommend caution in interpreting their results.
Issues in Analyzing Environmental Exposures and Preterm Birth
A major issue in studying environmental exposures and preterm birth is selecting the relevant
exposure period, since the biological mechanisms leading to preterm birth and the critical periods of
vulnerability are poorly understood (Bobak, 2000). Exposures proximate to the birth may be most
relevant if exposure causes an acute effect. However, exposure occurring in early gestation might affect
placentation, with results observable later in pregnancy, or cumulative exposure during pregnancy may be
the most important determinate. The studies reviewed have dealt with this issue in different ways. Many
have considered several exposure metrics based on different periods of exposure.
Often the time periods used are the first month (or first trimester) of pregnancy and the last month
(or six weeks) prior to delivery. Using a time interval prior to delivery introduces an additional problem
since cases and controls are not in the same stage of development when they are compared. For example,
a preterm infant delivered at 36 weeks, is a 32 week fetus four weeks prior to birth, while an infant born
at term (40 weeks) is a 36 week fetus four weeks prior to birth. Only one study (Huynh et al., 2006)
adjusted for this in the design.
Many of these studies compare exposure in quartiles, using the lowest quartile as the reference (or
control) group. No studies use a truly unexposed control group. If exposure in the lowest quartile confers
risk, than it may be difficult to demonstrate additional risk associated with a higher quartile. Thus
negative studies must be interpreted with caution.
Preterm birth occurs both naturally (idiopathic preterm), and as a result of medical intervention
(iatrogenic preterm). Ritz et al. (2000; 2007) excluded all births by Cesarean section, to limit their studies
to idiopathic preterm. No other studies attempted to distinguish the type of preterm birth, although
exposure maybe associated with only one type. This is another source of potential misclassification.
December 2008
7-61
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Growth Restriction
Low birth weight has often been used as an outcome measure because it is easily available and
accurately recorded on birth certificates. However, low birth weight may result from either short
gestation, or inadequate growth in utero. Most of the studies investigating air pollution exposure and low
birth weight, limited their analysis to term infants to focus on inadequate growth. Five studies were
identified that specifically addressed growth restriction in utero by identifying infants who failed to meet
specific growth standards. Usually these infants had birth weights less than the 10th percentile for
gestational age, using an external standard. All five of these studies have been previously discussed, since
they also examined other reproductive outcomes (low birth weight or preterm delivery).
Two studies in the U.S. examined intrauterine growth and both were conducted in California.
Parker et al. (2005) reported a positive association between exposure to PM2 5. Since this study only
included singleton live births at 40 weeks gestation, birth weights less than 2872 grams for girls and 2986
grams for boys were designated SGA, based on births in California. Infants exposed to the highest
quartile PM25 (>18.4 (.ig/nr3) compared to the lowest quartile PM25 (<11.9 (.ig/nr3) were 23% more likely
to be small for gestational age (OR = 1.23 [95% CP 1.03-1.50]). Very similar results were found for
exposure in each of the three trimesters respectively (OR= 1.26 [95% CP 1.04-1.51], OR= 1.24 [95%
CP 1.04-1.49], OR =1.21 [95% CP 1.02-1.43]). These results controlled for exposure to CO, which did
not increase risk for SGA.
In contrast, Salam(2005) found no association between exposure to PM10 and intrauterine growth
retardation (IUGR) in the California Children's Health Study. IUGR was defined as less than the 15th
percentile of predicted birth weight based on gestational age and sex in term infants. Apparently no
external standard was used since 15% of infants in the study were designated as IUGR. An IQR increase
in PM10 exposure was not significantly associated with IUGR for the whole pregnancy (OR = 1.1 [95%
CI: 0.9-1.3]) or for any specific trimester. Differences between this study and the study by Parker et al.
(2005) include measurement of PMi0 vs. PM2 5, a less stringent definition of IUGR, and exposures
determined by monitors located much farther away from the subjects' residences (up to 50 km vs. within
5 miles). All of these factors could lead to misclassification.
Two studies investigating particle exposure and SGA were conducted in Australia, with differing
results (Hansen et al., 2007a; Mannes et al., 2005). Mannes et al. (2005) defined SGA as birth weight less
than two standard deviations below the national mean birth weight for gestational age. In this study there
was a statistically significant effect of exposure to both PMi0 (OR =1.10 [95% CI: 1.00-1.48], per
10 |ig/m3 increase) and PM25 (OR = 1.34 [95% CI: 1.10-1.63], per 10 |ig/m3 increase) for exposure during
the second trimester. When analysis was restricted to births within 5 km of the monitoring station, the
December 2008
7-62
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
association for PM10 became slightly stronger (OR = 1.22 [95% CI: 1.10-1.34]). Exposure during other
trimesters of pregnancy was not associated with IUGR.
In Brisbane, Hansen et al. (2007a) examined head circumference (HC), crown heel length (CHL)
and risk of SGA, defined as less than the tenth percentile of weight for gestational age and gender based
on an Australian national standard. There was no consistent relationship between PMi0 exposure and
SGA, HC or HCL in any trimester of pregnancy. PMi0 exposure was determined by averaging values
from the five monitoring stations. Due to the sample size and limited number of monitoring stations, it
was not possible to analyze the data for women living within 5 km of a monitoring station, as was done in
Sydney.
In Canada, Liu (2007d) investigated the effect of PM25 exposure on fetal growth in three cities,
Calgary, Edmonton and Montreal. Intrauterine growth retardation (IUGR) was defined as birth weight
below the tenth percentile, by sex and gestational week (37-42) for all singleton live births in Canada
between 1986 and 2000. Models were adjusted for maternal age, parity, infant sex, season of birth, city of
residence, and year of birth. A 10 (ig/m3 increase in PM2.5 was associated with an increased risk for IUGR
(OR= 1.07 [95% CI: 1.03-1.10]) in the first trimester, and similar risks were associated with exposure in
the second or third trimesters. The effect of PM2.5 was reduced in multipollutant models including CO and
N02.
Brauer et al. (2008) observed consistent increased risks of SGA for PM2 5, PM10, N02, NO and CO
in Vancouver, Canada (20% increase in risk in PM2 5 and PM10 per 10 (.ig/ni3 increase). The effects were
similar for exposure estimates based on nearest monitor, inverse distance weighting, and land-use
regression modeling. ORs for early or late pregnancy exposure windows were remarkably similar to those
for the full duration of pregnancy.
One additional study investigating fetal growth was conducted in the Czech Republic (Bobak,
2000). There was no association between total suspended particulate (TSP) exposure and IUGR (defined
as less than the tenth percentile of weight for gestational age and gender). For example the odds ratio for a
50 ng/m3 increase in TSP was OR = 0.89 [95% CI: 0.75-1.06],
Birth Defects
Three recent articles examined PM and birth defects. The Seoul, Korea study mentioned above also
considered congenital anomalies, defined as a defect in the child's body structure (Kim et al., 2007a).
PMio levels were associated with higher risk of birth defects for the second trimester, with a 16% (95%
CI: 0-34) increase in risk per 10 (ig/m3 in PMi0.
Two U.S. studies specifically examined air pollution and risk of birth defects. Data were collected
from the California Birth Defects Monitoring Program for four counties in Southern California (Los
December 2008
7-63
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Angeles, Riverside, San Bernardino, and Orange) for the period 1987 to 1993, although each county
included a subset of this period (Ritz et al., 2002). Cases (i.e., infants with birth defects) were identified as
live birth infants and fetal deaths from 20 weeks gestation to 1 year post birth, with isolated, multiple,
syndrome, or chromosomal cardiac or orofacial cleft defects. Cases were restricted to those with registry
data for gestational age and residence zip code, and those with residences <10 miles from an air pollution
monitor. Six types of categories were included: aortic defects; atrium and atrium septum defects;
endocrinal and mitral value defects; pulmonary artery and valve defects; conotruncal defects; and
ventricular septal defects not part of the conotruncal category. PMi0 measurements were available every
six days. While results indicated increased risk of birth defects for higher levels of CO or 03, the authors
determined that results for PMi0 were inconclusive, finding no consistent trend of effect after adjustment
for CO and 03.
The other U.S. study examined birth defects through a case-control design in seven Texas counties
for the period 1997 to 2000 (Gilboa et al., 2005). Births were excluded for parents <18 years and several
non-air pollution risk factors known to be associated with birth defects (e.g., maternal diabetes,
holoprosencephaly in addition to oral clef). Comparison of the highest (>29.0 (.ig/nr1) and lowest
(<19.521 (.ig/nr1) quartiles of PMi0 for exposure defined as the third to eighth weeks of pregnancy
generated an OR of 2.27 (95% CI: 1.43-3.60) for risk of isolated artrial septal defects and 1.26 (95% CI:
1.03-1.55) for individual artrial septal defects. Including other pollutants (CO, N02, 03, S02) in the
model did not greatly alter results; numerical results for copollutant analysis were not provided. Strong
evidence was not observed for a relationship between PM10 and the other birth defect categories. Review
articles have concluded that the scientific literature is not sufficient to conclude a relationship between air
pollution and birth defects (Sram et al., 2005).
Infant Mortality
Many studies have identified strong associations between exposure to particles and increased risk
of mortality in adults or the general population, including for short- and long-term exposure (Pope and
Dockery, 2006; U.S. EPA, 2004). Less evidence is available for the potential impact on infant mortality,
although studies have been conducted in several countries. The results of these infant mortality studies are
presented here with the other reproductive and developmental outcomes because it is likely that in vitro
exposures contribute to this outcome. Both long-term and short-term exposure studies of infant mortality
are included in this section. Results on PM and infant mortality includes a range of findings, with some
studies finding associations and many non-statistically significant or null effects. Yet, more consistency is
observed when results are divided into the type of health outcome based on the age of infant and cause of
death.
December 2008
7-64
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
An important question regarding the association between PM and infant mortality is the critical
window of exposure during development for which infants are susceptible. Several age structures have
been explored: infants (<1 year); neonatal (<1 month); and postneonatal (1 month to 1 year). Within these
various age categories, multiple causes of deaths have been investigated, particularly total deaths and
respiratory-related deaths. The studies reflect a variety of study designs, particle size ranges, exposure
periods, regions, and adjustment for confounders.
Stillbirth
Only one study of stillbirths and PM published from 2002 to present was identified. A prospective
cohort of pregnant women in Seoul, Korea from 2001 to 2004 was examined with respect to exposure to
PMio (Kim et al., 2007a). Gestational age was estimated by the last menstrual period or by ultrasound.
Whereas many of the previously discussed studies of PM and pregnancy outcomes were based on national
registries, this study examined medical records and gathered individual information through interviews on
socio-economic condition, medical history, pregnancy complications, smoking, second-hand smoke
exposure, and alcohol use. Mother's exposure to PM10 was based on residence for each month of
pregnancy, each trimester defined as a three month period, and the six weeks prior to death. Exposure was
assigned by the nearest monitor. A 10 (ig/m3 increase in PMi0 in the third trimester was associated with an
8% (95% CI: 2-14) increase in risk of stillbirth.
Earlier research investigating PM and stillbirths includes an ecological study of the Czech Republic
for the period 1986 to 1988 based on the frequency of stillbirths and TSP levels (Bobak and Leon, 1999a).
Risk of stillbirths, defined as a deceased infant >1000 g or > 28 weeks gestation, was not associated with
TSP in this research. In Sao Paulo, Brazil, Poisson regression of stillbirth counts for the period 1991 and
1992 found that a 10 (ig/m3 increase in PMi0 was associated with a 0.8% increase in stillbirth rates
(Pereira et al., 1998). When other pollutants (N02, S02, CO, 03) were included simultaneously in the
model, the association did not remain. Stillbirths were defined as fetal loss at >28 weeks of pregnancy
age, weight >1000 g, or length of fetus >35 cm.
In summary, while there exists some evidence for a link between PM and stillbirths, the scientific
literature is not sufficient to draw this definitive conclusion. A review article concluded that there is
insufficient evidence to determine an effect of PM exposure and risk of stillbirth (Heinrich and Slama,
2007).
Infant Mortality and Infant Respiratory Mortality, <1 year
A literature search did not reveal new studies on PM and infant mortality (<1 year) since the
previous PM AQCD. Previously conducted studies include a case-control study that reported associations
December 2008
7-65
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
between infant mortality and TSP levels over the period between birth and death for infants in the Czech
Republic (Bobak and Leon, 1999a); a 50 (ig/m3 increase in TSP was associated with a 77% (95% CI:
5-301) increase in respiratory-related infant mortality. An ecological study evaluated U.S. PM10 data for
the year 1990 using long-term pollution levels in 180 U.S. counties (Lipfert et al., 2000). The authors
found a 9.64% (95% CI: 4.60-14.9) increase in risk of infant mortality for non-low birth weight infants
per 10 |_ig/m3 increase in PMi0, a 13.4% (95% CI: -10.3 to 43.5%) increase in non-low birth weight
respiratory-disease related deaths (ICD-9 460-519) and a 19.5% (95% CI: 0.07-42.8) increase in all non-
low birth weight respiratory-related infant deaths (ICD 9 460-519, 769, 770).
Neonatal Mortality and Neonatal Respiratory Mortality, <1 month
Studies on PM and neonatal mortality (<1 month) included a time-series analysis of PMi0 for four
years of data (1998-2000) for Sao Paulo, Brazil (Lin et al., 2004a). The analysis used daily counts of
deaths from government registries and adjusted for temporal trend, day of the week, weather, and
holidays. Findings indicated that a 10 (ig/m3 increase in PM10 was associated with a 1.71% (95% CI:
0.31-3.32) increase in risk of neonatal death.
A case-crossover study of 11 years (1989-2000) in Southern California did not find an association
between PMi0 and neonatal deaths (Ritz et al., 2006). Numerical results were not provided. The authors
considered adjustment for season, county, parity, gender, prenatal care, and maternal age, education, and
race/ethnicity. The overall levels of PMi0 in these studies were similar.
These results add to previous work on PM and neonatal death, including studies identifying higher
risk of neonatal mortality with higher TSP in the Czech Republic in an ecological analysis (Bobak and
Leon, 1992) and case-crossover study (Bobak and Leon, 1999a), and a Poisson model study in
Kagoshima City, Japan (Shinkura et al., 1999). An ecological study evaluated U.S. PMi0 data for the year
1990 using long-term pollution levels in 180 U.S. counties (Lipfert et al., 2000). Analysis considered birth
weight, sex, month of birth, location by state and county, prenatal care, and mother's race, age,
educational level, marital status, and smoking status. County-level variables were included for socio-
economic status, altitude, and climate. Results indicate a 13.1% increase in neonatal mortality (95% CI:
4.4-22.6) per 10 (ig/m3 PMi0 for non-low birth weight infants. Statistically significant associations were
also observed considering all infants or low birth weight infants. However, higher levels of S02 were
associated with lower risk of infant mortality. When sulfate and an estimate of non-sulfate particles were
included in the regression simultaneously, associations were observed with non-sulfate particles and an
inverse relationship with sulfate particles. Respiratory neonatal mortality was not associated with higher
TSP in the Czech Republic case-control study (Bobak and Leon, 1999a).
December 2008
7-66
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Postneonatal Mortality and Post-neonatal Respiratory Mortality, 1 month-1 year
Several studies have been conducted on PM and postneonatal mortality since the previous PM
AQCD, including three from the U.S., one from Mexico, and three from Asia. Two case-control studies
examined the risk of PM to postneonatal death in California. Research focused on Southern California for
the period 1989 to 2000 linked birth and death certificates and considered PMi0 two months prior to death
with adjustment for prenatal care, gender, parity, county, season, and mother's age, race/ethnicity, and
education (Ritz et al., 2006). As previously noted, this study did not find an association between PMi0 and
neonatal mortality (<1 month), however an association was observed for postneonatal mortality, with a
10 (ig/m3 increase in PMi0 associated with a 4% (95% CI: 1-6) increase in risk. The exposure period of
two weeks before death was also considered, producing effect estimates of 5% (95% CI: 1-10) for the
same PM10 increment. Even stronger estimates were observed for those who died at ages 4 to 12 months.
When CO, N02, and 03 were simultaneously included with PM10 in the model, the central estimate
reduced to 2% for the 2-week exposure period and 4% for the 2-month exposure period, and both
estimates lost statistical significance. The other case-control study of California considered PM2 5 from
1999 to 2000 for infants born to mothers within five miles of a PM2 5 monitoring station (Woodruff et al.,
2006). Infants who died during the postneonatal period were matched to infants with date of birth within
two weeks and birth weight category. Exposure was estimated from the time of birth to death. Models
considered parity and maternal race, education, age, and martial status. A 10 (ig/m3 increase in PM2.5 was
associated with a 7% (95% CI: -7 to 24) increase in postneonatal death
County-level PMi0 and PM2 5 for the first two months of life for births in urban U.S. counties (>
250,000 residents) from 1999 to 2002 were evaluated in relation to postneonatal mortality with GEE
models (Woodruff et al., 2008). Analyses were adjusted for primiparity (first born), community-level
poverty, region, month, year, and mother's race, marital status, education, and age. Births were restricted
to singleton births with gestational age < 44 weeks, same county of residence at birth and death, and non-
missing data on birth order, birth weight, and maternal race, education, and martial status. Higher levels
of either PM metric were associated with higher risk of postneonatal mortality, with 4% (95% CI: -1 to
10) increase in mortality risk per 10 (ig/m3 in PMi0 and 4% (95% CI: -2 to 11) increase in mortality risk
for the same increment of PM2 5. This work builds on a previous study of 86 U.S. urban areas from 1989
to 1991, finding a 4% (95% CI: 2 to 7) increase in postneonatal mortality per 10 (ig/m3 county-level PM10
over the first two months of life (Woodruff et al., 1997).
In Ciudad Juarez, Mexico, a case-crossover approach was applied to data from 1997 to 2001 based
on death certificates and the cumulative PM10 for the day of death and previous two days (Romieu et al.,
2004). A case-crossover study of Kaohsiung, Taiwan from 1994 to 2000 compared the average of PM10 on
the day of death and two previous days to PMi0 in control periods a week before and week after death
December 2008
7-67
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
(Tsai et al., 2006b). A similar approach was also applied to 1994 to 2000 data from Taipei, Taiwan, also
using case-crossover methods for the lag 0-2 PM10 with referent periods the week before and after death
(Yang et al., 2006). In these case-crossover studies, season was addressed through matching in the study
design. A 10 (ig/m3 increase in PMi0 was associated with a 2.0% (95% CI: -2.8 to 7.0) increase in the
Mexico study, a 0.59 (95% CI: -15.0 to 18.8) increase in postneonatal death in the Kaohsiung study, and a
1.02% (95% CI: -13.2 to 17.6) increase in the Taipei study. A study in Seoul, South Korea from 1995 to
1999 used time-series approaches adjusted for temporal trend and weather, based on national death
registries excluding accidental deaths (Ha et al., 2003). A 10 |_ig/m3 increase in PMi0 was associated with
a 3.14% (95% CI: 2.16 to 4.14) increase in risk of death for postneonates.
These studies add to evidence from two earlier studies in the Czech Republic, described above,
which found higher risk of postneonatal mortality with higher PM levels. The association was uncertain in
a case-control study based in the Czech Republic (Bobak and Leon, 1999a). In an ecological study of 45
Czech Republic districts, a statistically significant trend was observed with higher risk of postneonatal
mortality at higher quintiles of TSP-10 (TSP up to 10 |_im) (Bobak and Leon, 1992).
A subset of the studies examining postneonatal mortality also considered the subset of postneonatal
deaths from respiratory causes. These include the time-series study in South Korea, finding a 17.8% (95%
CI: 14.4 to 21.2) increase in respiratory-mortality per 10 (ig/m3 increase in PM10 (Ha et al., 2003) and the
case-crossover study in Mexico, for which the same increment in PM10 was associated with a 1.5% (95%
CI: -14.1 to 13.0) decrease in risk (Romieu et al., 2004). Both case-control California studies identified
associations, with a 5% (1, 10%) increase in risk in Southern California (Ritz et al., 2006) and 57.4%
(95% CI: 7.0 to 132) increase in California per 10 (ig/m3 PM10 (Woodruff et al., 2006). The U.S. study
found this increment in PMi0 to be linked with a 16% (95% CI: 6.0 to 28.0) increase in respiratory
postneonatal mortality, although effect estimates for PM2 5 were not statistically significant (Woodruff et
al., 2008). Studies conducted prior to 2002 on respiratory-related postneonatal mortality include both
Czech Republic studies and the study of 86 U.S. urban areas, all finding statistically significant effects
(Bobak and Leon, 1992, 1999a; Woodruff et al., 1997).
Sudden Infant Death Syndrome
Three studies examining the relationship between PM and sudden infant death syndrome have been
published from 2002 onward. These studies examined infant mortality and were thereby discussed in this
section previously. A case-control study over a 12-year period (1989 to 2000) matched 10 controls to
deaths (cases) in Southern California (Ritz et al., 2006). A 10 (ig/m3 increase in PM10 the two months
prior to death was associated with a 3% (95% CI: -1 to 8) increased in SIDS. Adjusted for other pollutants
(CO, N02, and 03), the effect estimate reduced to 1% (95% CI: -5 to 7).
December 2008
7-68
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
A case-control study, also based in California, found an OR of 1.008 (95% CI: 1.006-1.012) per
10 (ig/m3 increase in PM25, considering a SIDS definition of ICD 10 R95(Woodruff et al., 2006). Due to
changes in SIDS diagnosis, another SIDS definition was explored for ICD 10 R99 in addition to ICD 10
R95. Under this SIDS definition, the effect estimate changed to 1.03 (95% CI: 0.79-1.35). The authors
also examined whether the relationship between PM2 5 and SIDS differed by season, finding no significant
difference. PMi0 and PMi0.2.5 were not associated with risk of SIDS; numerical results were not provided
for these PM metrics. The third recent study of PM and SIDS examined U.S. urban counties from 1999 to
2002 (Woodruff et al., 2008). Non-statistically significant relationships were observed between SIDS and
PMio or PM2 5 in the first two months of life.
These studies add to earlier work, such as a U.S. study that found higher risk of SIDS with higher
annual PM2 5 levels, including in a separate analysis of normal birth weight infants (Lipfert et al., 2000),
and a U.S. study identifying a 12% (95% CI: 7-17) increase in SIDS risk per 10 (ig/m3 in PMi0 for the
first two months of life for normal weight births (Woodruff et al., 1997). A study based on Taiwan found
higher SIDS risk with lower visibility (Knobel et al., 1995), whereas a 12 city Canadian time-series study
identified no significant associations (Dales et al., 2004).
Deaths by SIDS were identified by different methods in the studies, partly due to transition from
ICD 9 to ICD 10, but also due to different choices within the research design. Two studies examined
multiple approaches (ICD 10 R95, ICD 10 R95 and R99) (Woodruff et al., 2006; 2008), and other studies
investigated ICD 9 798.0 and ICD 10 R95 (Ritz and Wilhelm, 2008), ICD 9 798.0 (Woodruff et al., 1997),
ICD 9 798.0 and 799.0 (Knobel et al., 1995), as well as a sudden unexplained death of infant <1 year for
which an autopsy did not identify a specific cause of death (Dales et al., 2004). These variations in the
definition of health outcomes add to differences in populations and study designs.
While some findings indicate a potential effect of PM on risk of SIDS, with the strongest evidence
perhaps from the case-control study in California (Woodruff et al., 2006), others do not find an effect or
observe an uncertain association. For the relationship between PM and SIDS, a 2004 review article
concluded consistent evidence exists compared to evidence for other infant mortality effects (Glinianaia
et al., 2004), whereas other reviews found weaker or insufficient evidence (Heinrich and Slama, 2007).
Another review concluded that the scientific literature on air pollution and SIDS suggests an effect, but
that further research is needed to draw a conclusion (Tong and Colditz, 2004).
Comparisons Across Studies and Key issues
Comparison of results across studies can be challenging due to several issues, including differences
in methodologies, populations and study areas, pollution levels, and the exposure timeframes used. Given
the large variation in study designs, the methods to address potential confounders vary. For example,
December 2008
7-69
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
weather and season were addressed in the case-control studies by matching, in the time-series study
through non-linear functions of temperature and temporal trend, and in the ecological study through
county-level variables. All studies included consideration of seasonality and weather, with the exception
of an older study based on the Czech Republic (Bobak and Leon, 1992). Researchers used different
definitions of respiratory-related deaths, including ICD 9 460-519 (Bobak and Leon, 1999a; Lipfert et al.,
2000); ICD 9 460-519, 769-770 (Lipfert et al., 2000); ICD 9 codes 460-519, 769, 770.4, 770.7, 770.8,
770.9, and ICD 10 J00-J98, P22.0, P22.9, P27.1, P27.9, P28.0, P28.4, P28.5, and P28.9 (Ritz et al., 2006);
and ICD 9 460-519 and ICD 10 J00-J99 for any cause on death certificate (Romieu et al., 2004); ICD 10
J00-99 and P27.1 excluding J69.0 (Woodruff et al., 2006; 2008); and ICD 9 460-519 (Woodruff et al.,
1997).
Socio-economic conditions were included at the individual level, typically through maternal
education, in many studies (e.g., Bobak and Leon, 1999a; Ritz et al., 2006; Woodruff et al., 1997;
Woodruff et al., 2006; 2008) and at the community-level in others (e.g., Bobak and Leon, 1992; Penna
and Duchiade, 1991), or for both individual and community-level data (e.g., Lipfert et al., 2000). The
time-series approach is unlikely to be confounded by socio-economic and other variables that do not
exhibit day-to-day variation. Similarly, case-crossover methods use each case as his/her own control,
thereby negating the need for individual-level confounders such as socio-economic status (e.g., Romieu et
al., 2004; Tsai et al., 2006b; Yang et al., 2006). All studies published after 2001 incorporated individual-
level socio-economic data or were of case-crossover or time-series design. One study specifically
examined whether socio-economic status modified the PM and mortality relationship, dividing subjects
into three socio-economic strata based on the zip code of residence at death (Romieu et al., 2004). This
work, based in Mexico, found that at lower socio-economic levels the association between PMi0 and
postneonatal mortality increased. While the overall association showed higher risk of death with higher
PMio with statistical uncertainty, for the lowest socio-economic group, a 10 (ig/m3 increment in
cumulative PMi0 over the two days before death was associated with a 60% (95% CI: 3-149) increase in
post-neonatal death. A trend of higher effect for lower socio-economic condition is observed in all three
lag structures.
Studies differ in terms of the timeframe of pregnancy that was used to estimate exposure. Exposure
to PM for infant mortality (<1 year) was estimated as the levels between birth and death (Bobak and
Leon, 1999a), annual community levels (Lipfert et al., 2000; Penna and Duchiade, 1991), and the 3 to 5
days prior to death (Loomis et al., 1999). For neonatal deaths, exposure timeframes considered were the
time between birth and death (Bobak and Leon, 1992, 1999a), annual levels (Bobak and Leon, 1999a;
Lipfert et al., 2000), monthly levels (Shinkura et al., 1999), the same day concentrations (Lin et al.,
2004a), and the two months or two weeks prior to death (Ritz et al., 2006). Postneonatal mortality was
December 2008
7-70
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
examined with PM concentrations based on annual levels (Bobak and Leon, 1992; Lipfert et al., 2000),
between birth and death (Bobak and Leon, 1999a; Woodruff et al., 2006), two months before death (Ritz
et al., 2006), the first two months of life (Woodruff et al., 1997; 2006), the day of death (Ha et al., 2003),
and the average of the same day as death and previous two days (Romieu et al., 2004; Tsai et al., 2006b;
Yang et al., 2006). Thus, no consistent window of exposure was identified across the studies.
Pollution levels were highest in the Czech Republic (68.5 |_ig/m3) (Bobak and Leon, 1992), South
Korea (69.2 (.ig/nr3) (Ha et al., 2003), and Taiwan (81.45 (.ig/m3) (Tsai et al., 2006b), and lowest in the
U.S. (29.1 |ag/m3) (Woodruff et al., 2008) and Japan (21.6 |_ig/m3) (Shinkura et al., 1999). All studies used
community-level exposure information based on ambient monitors, as opposed to exposure measured at
the individual level (e.g., subject's home) or personal monitoring.
Given similar sources for multiple pollutants (e.g., traffic), disentangling the health responses of
copollutants is a challenge in the study of ambient air pollution. Several studies examined multiple
pollutants, most by estimating the effect of different pollutants through several univariate models. Some
studies noted the difficulty of separating particle impacts from those of other pollutants, but noted
stronger evidence for particles than other pollutants (Bobak and Leon, 1999a). A few studies applied
copollutant models by including multiple pollutants simultaneously in the same model. Effect estimates
for the relationship between PM10 and neonatal deaths in Sao Paulo were reduced to a null effect when
S02 was incorporated (Lin et al., 2004a). Associations between PM10 and postneonatal mortality or
respiratory postneonatal mortality remained but lost statistical significance in a multiple pollutant model
with CO, N02, and 03 (Ritz et al., 2006).
Several review articles in recent years have examined whether exposure to PM affects risk of infant
mortality, generally concluding that more consistent evidence has been observed for postneonatal
mortality, particularly from respiratory causes (Glinianaia et al., 2004; Heinrich and Slama, 2007;
Lacasana et al., 2005; Sram et al., 2005). In one review authors identified 14 studies on infant mortality
and air pollution and determined that studies on PM and infant mortality do not provide consistent results,
although more evidence was present for an association for some subsets of infant mortality such as
postneonatal respiratory-related mortality (Glinianaia et al., 2004). The relationship between PM and
postneonatal respiratory mortality was concluded to be causal in one review (Sram et al., 2005), and
strong and consistent in another (Heinrich and Slama, 2007). Meta-analysis using inverse-variance
weighting of PMi0 studies found that a 10 (ig/m3 increase in acute PMi0 exposure was associated with
3.3% (95% CI: 2.4-4.3) increase in risk of postneonatal mortality, whereas the same increment of chronic
PMio exposure was linked with a 4.8% (95% CI: 2.2-7.2) increase in postneonatal mortality and a 21.6%
(95% CI: 10.2-34.2) increase for respiratory postneonatal mortality (Lacasana et al., 2005). This review
noted that "The studies on infant mortality and exposure to particles show an outstanding consistency in
December 2008
7-71
DRAFT-DO NOT QUOTE OR CITE
-------
1 the magnitude of the effects, regardless of the different designs used" (Lacasana et al., 2005). Another
2 meta-analysis of five studies found a 10 |ag/nr increase in PM10 to be associated with a 5.6% (95% CI:
3 2.6-8.8) increase in infant mortality (Roosli et al., 2005). Other reviews note that there exists
4 "considerable evidence" that maternal exposure to PM is linked with adverse pregnancy health responses
5 (Schwartz, 2004a).
o
^r
o
CN
O
0
20
40
60
o
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
found higher effects for risk in the neonatal period. Another study, based in Southern California identified
no association for neonatal effects (numerical results not provided) but statistically significant results for
postneonatal mortaliy (Ritz et al., 2006). Figure 7.6 compares risk for the postneonatal period for
respiratory and total mortality. In six of the seven studies, higher effect estimates were observed for
respiratory-related mortality. Results from the neonatal period found higher effects for total mortality
compared to respiratory mortality (Bobak and Leon, 1999a) and the reverse for a study examining infant
mortality (Lipfert et al., 2000). Thus, there exists evidence for a stronger effect at the postneonatal period
and for respiratory-related mortality, although this trend is not consistent across all studies.
Decrements in Sperm Quality
Limited research conducted in the Czech Republic on the effect of ambient air pollution on sperm
production has found associations between elevated air pollution and decrements in proportionately fewer
motile sperm, proportionately fewer sperm with normal morphology or normal head shape,
proportionately more sperm with abnormal chromatin (Selevan et al., 2000), and an increase in the
percentage of sperm with DNA fragmentation (Rubes et al., 2005). These results were not specific to PM,
but for exposure to a high-, medium- or low-polluted air mixture. Research in Los Angeles, California
examined 5134 semen samples from 48 donors in relation to ambient air pollution measured 0-9, 10-14,
70-90 days before semen collection over a two year period (1996-1998). Ambient ozone during all
exposure periods had a significant negative correlation with average sperm concentration, and no other
pollutant measures were significantly associated with sperm quality parameters, or presented
quantitatively (Sokol et al., 2006).
7.4.2. Toxicological Studies
This section summarizes recent evidence on reproductive health effects reported with exposure to
ambient PM; no evidence was available in this area in the 2004 PM AQCD. Studies from different
toxicological rodent models allow investigation of specific mechanisms and modes of action for
reproductive changes. Emphasis is placed here on results from different windows of development, i.e., if
exposure in utero, neonatally or as an adult can lead to similar reproductive outcomes as an adult. In
addition, studies evaluating whether fertility is affected in female and/or male animals equally by a
similar exposure, and how exposures are transmitted to the fertility of the F1 offspring, are summarized.
Hormonal changes which can lead to decreased sperm count or changes in the estrous cycle are also of
interest. Pregnancy losses and placental sufficiency are also followedT. Most recently, the role of
December 2008
7-73
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
environmental chemicals in shifting sex ratios (also seen in epidemiologic studies) and in affecting
heritable DNA changes have become endpoints of interest.
Female Reproductive Effects
Significant work has been done in male rodent models to determine the effect of PM exposure on
reproductive success; fewer studies have been done on female rodents. Tsukue et al. (2004) exposed
pregnant C57-BL mice to DE 0.1 mg DEP/m3 diluted in charcoal and HEPA-filtered clean air or to for 8
h/day GD 2-13 and at GD 14 collected the female fetuses for analysis of mRNA for Ad4BP-l/SF-l and
MIS, and found no significant changes. The concentration of the gaseous materials including NO, NOx,
N02, CO and S02 are 2.2 + 0.34 ppm, 2.5 + 0.34 ppm, 0.0 ppm, 9.8 +/- 0.69 ppm, and <0.1ppm (not
detectable), respectively. Work by Yoshida et al. (2006b) showed changes in these two transcripts in male
ICR fetuses exposed to similar doses of DEP, albeit with different daily durations of exposure. Further
work by Yoshida et al. showed that of three mouse strains tested, ICR male fetuses were the most
sensitive to DE-dependent changes in these two genes. Nonetheless, strain sensitivity to DEP may also
differ by sex. Tsukue et al. (2004) also looked at mRNA transcript levels of other steriodogenic and
oocyte-affecting genes [aromatase, Cytochrome P450 1A1 (CYP 1A1), MIS, x-chromosome gene-1
(DAX-1), estrogen receptor (ER), Wingless-4 (Wnt), Wnt-7a, Growth differentiation factor-9 (GDF-9),
bone morphogenetic protein-6 (BMP-6) and BMP-15], Among these genes, they found significant mRNA
reduction of BMP-15, a gene related to oocyte development. Thus, it appears that female mice exposed in
utero to DE show a lack of response at the mRNA level of MIS or Ad4bP-l/SF-l, important genes in
male sexual differentiation that showed DE-dependent changes in male pups from dams exposed in utero.
Female fetuses do however show a decrease in BMP-15, which is related to oocyte development. Possible
ingestion exposure from grooming cannot be ruled out in this study.
Windows of exposure are important in determining reproductive success as an adult. Exposure as a
neonate may have a drastically more profound impact than does a similar adult exposure in females. To
test this, female BALB/C mice were exposed to ambient air in Sao Paulo as neonates or as adults and then
were bred to non-exposed males (Mohallem et al., 2005). Concentrations of pollutants in this ambient air
including CO, N02, PMi0, and S02 as measured locally were 2.2 ± 1.0, 107.8 ± 42.3, 35.5 ± 12.8, and
11.2 ± 5.3, respectively. They reported decreased fertility in newborn but not adult female BALB/c mice
after exposure to ambient air. There were a significantly higher number of liveborn pups from dams
housed in filtered chambers (PM removed as well as chemical substances) versus animals exposed to
ambient air as newborns. There was also a higher incidence of implantation failures in dams reared as
newborns in polluted chambers. Sex ratio (unlike in epidemiologic studies), number of pregnancies per
group, resorptions, fetal deaths, and fetal placental weights did not differ significantly by treatment group.
December 2008
7-74
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Thus, in these studies, exposure to ambient air pollution and its associated PM affected future
reproductive success of females if they were exposed as neonates and not if they are exposed as adults.
Environmental chemicals have been shown to act as endocrine disruptors by acting on the
androgen pathway, including the phthalates, which have manifested their anti-androgenic activity in
numerous pathways including decreased anogenital distance in male rodents (Foster et al., 1980; Foster et
al., 2001). To access the role of DE exposure on reproductive success and anti-androgenic effects on
offspring, Tsuke et al. (2002) exposed 6 week old female C57-BL mice to 4 months of DE (0.3, 1.0, or 3.0
mg/m3) or filtered air (controls). Some animals were euthanized at the end of this exposure; DE-exposed
estrous females from this group were found to have significantly decreased uterine weight (1.0 mg/m3).
Some of these DE-exposed females were bred to unexposed males. It was determined that DE-exposure
led to increased but not significant rates of pregnancy loss in mated females (up to 25%). The rate of good
nest construction of the pregnant exposed dams at the highest dose group was significantly lower than
control (3.0 mg DEP/m3). Offspring were weighed after birth with significant decreases in body weight
seen at 6 and 8 weeks (males and females 1.0 and 3.0 mg DEP/m3) and in female offspring (9 weeks of
age, 1.0 and 3.0 mg/m3). Anogenital distance, a sensitive marker of anti-androgen activity in males, was
significantly decreased in 30-day old DEP exposed male offspring (0.3 mg DEP/m3) v. controls. Thymus
weight was significantly decreased in 30-day old female offspring (3.0 mg DEP/m3) and remained
decreased at 70 days (0.3 and 1.0 mg DEP/m3). Ovary weight of female offspring was significantly
decreased (3.0 mg DEP/m3) at 30 days, but no longer significantly different at 70 days. In males at 70
days of age, body weights were significantly decreased and AGD was significantly shorter (3.0 mg
DEP/m3). In females at 70 days of age, the l.Omg DEP/m3 group showed significantly lower organ
weights (adrenals, liver, and thymus) and the 3.0 mg DEP/m3 group had decreased body weight. Thymus
weight of the 0.3 mg/m3 females was significantly lower at 70 days. Also, crown to rump length in
females from dams exposed to DEP (1.0 and 3.0 mg DEP/m3) was also significantly lower. In conclusion,
adult exposure to DEP led to maternal dependent reproductive changes that affected outcomes in
offspring manifesting as decreased pup body weight, anti-androgenic effects like decreased AGD and
decreased organ weight (which may be confounded by changes in body weight).
Male Reproductive Effects
Rodent strains differ in their sensitivity and response to various environmental chemicals. Studies
were performed to determine PM-dependent strain sensitivity using male steroidogenic enzymes as the
model pathway. In utero exposure of 3 strains of pregnant mice (ICR, C57B1/6J or ddY mice) via
inhalation exposure of DE at 0.1 mg DE particles (DEP)/m3 in HEPA-filtered clean air occurred or clean
air as controls continuously over gestational days 2-13 (Yoshida et al., 2006b). At gestational day 14,
December 2008
7-75
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
dams were euthanized and fetuses were collected from the uteri. Male fetuses were collected from each
dam for mRNA analysis of genes related to male gonad development including mullerian inhibiting
substance (MIS), steroid transgenic factor (Ad4BP/SF-l), an enzyme in the testosterone synthesis
pathway, cytochrome P450 cholesterol side chain cleavage enzyme (P450scc), and other steroidogenic
enzymes [17|3-hydroxysteroid dehydrogenase (HSD), cytochrome P450 17-a-hydroxylase (P450cl7), and
3-(3hydroxysteroid dehydrogenase (3(3HSD)]. There were significant decreases in MIS (ICR, and
C57BL/6 mice) and Ad4BP/SF-l (ICR mice) versus control at gestational day (GD) 14. SF-1
transcriptionally regulates T secretion. MIS is crucial in for sexual differentiation including mullerian duct
regression in males. The ddY strain showed no significant changes in Ad4BP/SF-l or MIS which the
authors hypothesized may be due to changes in 3(3-hD, which had marked changes in expression in the
ddY strain when compared to non-DE exposed controls. From these studies, it appears that mouse strains
with in utero exposure to DE show differential sensitivity in gonadal differentiation genes (mRNA) of
male offspring; ICR are the most sensitive, followed by C57BL/6 with ddY mice the least sensitive.
Yoshida et al. (2006a) also monitored changes in the male reproductive tract after in utero exposure
to DE. Timed-pregnant ICR dams were exposed during gestation (2dpc to 16dpc) to continuous DE
generated to concentrations of 0.3, 1.0 or 3.0 mg DEP/m3 in HEPA-filtered air or clean air as controls. PM
deposition on the fur and ingestion of the dams by grooming is another possible exposure route in this
study. The reproductive tracts of male offspring were monitored at 4 weeks postnatally. These pups
received possible exposure through lactation as dams exposed to DE during gestation nursed pups. There
was a threshold effect; 0.3 mg/m3 had no effect on male reproductive organ weight or serum testosterone
(T). Exposure to the higher doses (1.0 and 3.0 mg/m3) of DEP led to significant increases in reproductive
gland weight [testis, prostate, seminal vesicle (3.0 mg DEP/m3 only) and coagulating gland]. The
intermediate dose of 1.0 mg DEP/m3 induced significant increases in serum T. The organ weights are
presented as absolute numbers and not adjusted for body weight, which is sometimes problematic for
complete representation of hormonal changes as body weight may confound absolute organ weight
changes. Nonetheless, there were also significant decreases in mRNA for the steroidogenesis related
enzymes 3(3HSD (3.0 mg DEP/m3) and aromatase (3.0 mg DEP/m3). Transcripts relating to male sexual
differentiation [Mullerian inhibitory substance (mis) and steroid transgenic factor (AD4BP/SF-1), 1.0 and
3.0 mg DEP/m3] were also significantly decreased. Sexual differentiation is a tightly regulated process.
For example, SF-1 missense mutations result in XY individuals with external female genitalia. Thus the
effect of environmental DE-exposure should not be underscored.
This study demonstrated effects of DE exposure on male spermatogenesis. Exposure of pregnant
ICR mice to DE (2 dpc-16 dpc continuous inhalation exposure to 1.0 mg DEP/m3 in filtered air or to
filtered clean air) led to impaired spermatogenesis in offspring (Ono et al., 2007). Male offspring were
December 2008
7-76
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
followed at PND 8, 16, 21 (3 weeks), 35 (5 weeks) and 84 (12 weeks). After 16dpc but before termination
of the study, all of the animals were transferred to a regular animal care facility. No cross fostering was
performed in this experiment, so pups that were born to DE-exposed dams were also nursed on these
dams and may have received lactational exposure to DE from milk. The gaseous components of the
diluted DE included nitric oxide (NO), N02, sulfur dioxide (S02), and carbon dioxide (C02) at
concentrations of 11.75+/-1.18, 4.62+/-0.36, 0.21 +/- 0.01, and 4922 +/-244 ppm, respectively. The
average proportion of sulfur in the fuel during this study was 0.043%. Body weight was significantly
depressed at PNDs 8 and 35. Accessory gland relative weight was significantly increased at PND 8 and 16
only. Serum testosterone was significantly decreased at 3 weeks and at 12 weeks was significantly
increased. At 5 and 12 weeks, daily sperm production (DSP) was significantly decreased. FSH receptor
and star mRNA levels were significantly increased at 5 and 12 weeks, respectively. Relative testis weight
and relative epididymal weight were unchanged at all time points. All endpoints were measured at each
time point and if not mentioned above, those data reported no significant changes. Histological changes
showed Sertoli cells with partial vacuolization and a significant increase in testicular multinucleated giant
cells in the seminiferous tubules of DE exposed animals compared to control. This study indicates that in
utero exposure to DE had effects on spermatogenesis in offspring.
In utero exposure to DE and its effect on adult body weight, sex ratio, and male reproductive gland
weight was measured by Yoshida et al. (2006a). Pregnant ICR mice were exposed by inhalation to DE
(0.3, 1.0 or 3.0 mg DEP/m3 or to clean air) from 2dpc to 16dpc. Pups were allowed to nurse in clean air
on exposed dams until weaning and at PND 28, male pups were sacrificed. At this time, serum
testosterone and pup reproductive gland weight was determined. Significant increases in relative
reproductive organ weights were reported at 1.0 and 3.0 mg DEP/m3 for the seminal vesicle, testis,
epididymis, coagulating gland, prostate and liver. Male pup serum testosterone was significantly
increased at 1.0 mg DEP/m3. Mean testosterone positively correlated with testis weight, daily sperm
production, aromatase and steroidogenic enzyme message level (P450cc, cl7 lyase, and P450 aromatase).
Sex ratio did not differ in DE-exposed animals versus control. Male pup body weight of DE-exposed
animals was significantly increased at PND 28 (1.0 and 3.0 mg DEP/m3). These studies show that in utero
DE-exposure led to increased serum testosterone and increased reproductive gland weight in male
offspring early in life.
The effects of DE on murine adult male reproductive function were studied by exposing ICR male
mice (6 weeks of age) to DE (clean air control, 0.3, 1.0 or 3.0 mg DECP/m3) for 12 h/day for six months
with another group receiving a one month recovery of clean air exposure post-exposure (Yoshida and
Takedab, 2004). After six months exposure, there was a dose-dependent significant increase in
degeneration of seminiferous tubules of mice exposed to DEP. After six months, there was a significant
December 2008
7-77
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
decrease in daily sperm production (DSP)/g of testis tissue in DEP exposed animals. After six months
exposure to DEP plus one month recovery with clean air exposure, significant decreases remained in DSP
at the higher doses; the effect was lost at 0.3 mg/m3. This adult exposure and other work with in utero
exposure to DE showed similar outcomes. The effect of ingestion of deposited PM from the fur during
inhalation exposure cannot be ruled out as a possible mechanism of exposure in this experiment.
Earlier studies showed an inverse correlation between environmental levels of PM and sperm count
in adult men (Mehta and Anad Kumar, 1997). To expand on PM-dependent changes in spermatogenesis,
an eloquent DE-exposure model was designed to determine if PM or the gaseous phase of DE was
responsible for changes in sperm production in rodents (Watanabe, 2005). Pregnant dams (F344/DuCij
rats) exposed to DE (6 hours/day exposure to 0.17 or 1.71 mg DEP/m3) or filtered air (removing PM only,
high dose filtered air and low dose filtered air) from GD7 to parturition produced adult offspring with a
decreased number of Sertoli cells and decreased daily sperm production (PND 96) when compared to
control mice exposed to clean air (Watanabe, 2005). The concentrations of N02 for the high filtered and
low filtered exposure groups were 0.8 and 0.1 ppm, respectively. Because both filtered and DE-exposure
groups showed the same outcomes, the effects are likely due to gaseous components of DE.
Another source of PM emissions that is common around the world is motorcycle exhaust. Adult
male (8 week old) Wistar rats were exposed to motorcycle exhaust (ME) for 1-h in the morning and 1-h in
the afternoon Monday through Friday at 1:50 dilution in filtered clean air for 4 weeks (group A) or 1:10
for 2 (group B) or 4 weeks (group C) or to clean air (Huang et al., 2008) via a head and nose inhalation
chamber. After 4 weeks of exposure, both exposed groups had significantly decreased body weight v.
control. All three groups showed a decreased number of spermatids in the testis after ME exposure. Both
1:10 exposure groups also showed a decrease in caudal epididymal sperm counts. Group C showed
significant decreases in testicular weight. Group C had decreased mRNA for the cytochrome P450
substrate 7-ehtoxycoumarin O-de-ethylase, and increased IL-6, IL-1B, and cox-2 mRNA control.
Decreased protein levels of antioxidant superoxide dismutase and increased IL-6 protein were reported
for group C when compared to control. Serum testosterone was significantly decreased in group C and co-
treatment of group C with the antioxidant vitamin E resulted in partial rescue of serum T levels and
caudal epididymal sperm counts (albeit still significantly decreased versus control), and returned IL-6,
IL-ip, and COX-2 ME-exposure-dependent message levels to baseline. The glutathione antioxidant
system and lipid peroxidation were unchanged after these ME exposures at the time points measured.
Male animals exposed to ME in this experiment showed significant decrements in body weight, spermatid
number, and serum testosterone with an increase in inflammatory cytokines. Vitamin E co-treatment with
ME-exposure led to an attenuation of inflammation and a partial rescue of testosterone levels and sperm
December 2008
7-78
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
numbers. No filtration was done in this experiment to determine if the toxic phase of the exposure was
gas or particulate-dependent.
Developmental Effects
Sex Ratio
A direct correlation between air pollution (PMi0) exposure and a decrease in standardized sex ratios
(SSRs) has been reported in humans exposed to air pollution (Lichtenfels et al., 2007), with fewer
numbers of male births reported. To understand this shift, two groups (control and exposed) of male Swiss
mice were housed concurrently in Sao Paulo and received either ambient air exposure or filtered air
(chemical and particulate filtering) from PND10 for four months. Filtration efficiency for PM2 5, carbon
black, and N02 inside the chamber was found to be 55%, 100%, and 35%, respectively. After this
exposure, non-exposed females were placed in either chamber to mate. After mating, the males were
sacrificed and testes collected; males exposed to ambient air showed decreased testicular and epididymal
sperm counts, decreased total number of germ cells, and decreased elongated spermatids, but no
significant change in litter size. Females were housed in the chambers and sacrificed on GDI9 when the
number of pups born alive and the sex ratio were obtained. There was a significant decrease in the SSR
for pups born after living in the ambient air-exposed chamber compared to the filtered chamber. In this
study, a shift in SSR has been shown for both humans and rodents exposed to air pollution, but other
studies with DE exposure (Yoshida et al., 2006a) or ambient air in Sao Paulo (Mohallem et al., 2005)
showed no changes in rodent sex ratio. Possible exposure to PM and other components of ambient air via
ingestion during grooming cannot be ruled out in this rodent model.
Immunological Effect-Placenta
Placental insufficiency can lead to the loss of a pregnancy or to adverse fetal outcomes. DE-
exposure has been shown to induce inflammation in various models. Fujimoto et al. (2005) accessed
cytokine/immunological changes of DE-dependent inhalation exposure on the placenta during pregnancy.
Pregnant Sic: CR mice were exposed to DE (0.3, 1.0, or 3.0 mg DEP/m3 in HEPA and charcoal-filtered
clean air from 2dpc to 13 dpc) or clean air in inhalation chambers; dams, placenta, and pups were
collected at 14dpc. There was a significant increase in the number of absorbed placentas in DE exposed
animals (0.3 and 3.0 mg DEP/m3) with a significant decrease in the number of absorbed placentas in DE
exposed animals at the middle dose (1.0 mg DEP/m3). Absorbed placentas from DE exposed mice had
undetectable levels of CYP1A1 and two fold increases in TNF-a; CYP1A1 placental mRNA from healthy
placentas of DE-exposed mice was unchanged versus control. Interleukin (IL)-2, IL-5, IL-12a, IL-12-|3
and granulocyte macrophage colony-stimulating factor (GM-CSF) mRNA significantly increased in
December 2008
7-79
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
placentas of DE-exposed animals (0.3 and 3 mg DEP/m3). Placental IL-6 mRNA was increased ten-fold
in DE-exposed mice (3.0 mg DEP/m3). Fujimoto et al. reported DE-induced significant increases in
multiple inflammatory markers in the placenta with significant increases in the number of absorbed
placentas.
Immunological Effects: Asthma
In utero exposure may confer susceptibility to PM-induced asthmatic responses in offspring.
Exposure of pregnant BALB/c mice to aerosolized ROFA leachate by inhalation or to DEP intranasally
increases asthma susceptibility to their offspring (Fedulov et al., 2008; Hamada et al., 2007). The
offspring from dams exposed for 30 min to 50 mg/ml ROFA 1, 3, or 5 days prior to delivery responded to
OVA immunization and aerosol challenge with airway hyperreactivity and increased antigen-specific IgE
and IgGl antibodies. Airway hyperreactivity was also observed in the offspring of dams intra-nasally
instilled with 50 (ig of DEP or Ti02, or 250 |_ig CB, indicating that the same effect could be demonstrated
using relatively "inert" particles. Pregnant mice were particularly sensitive to exposure to DEP or Ti02
particles, and genetic analysis indicated differential expression of 80 genes in response to Ti02 in
pregnant dams. Thus pregnancy and in utero exposure may enhance responses to PM, and exposure to
even relatively inert particles may result in offspring predisposed to asthma.
Placental Weights and Birth Outcomes
Pregnant female Swiss mice were exposed to ambient air (Sao Paulo) or filtered air over various
portions of gestation to determine if there was an association between fetal or placental weight or birth
outcomes with exposure to air pollution. The concentration of various components of the ambient air as
measured by a State Environmental Sanitation Agency 100 meters away from the rodent exposure
chambers reported PMi0 (42 ± 17 |im/m3). N02 (97 ± 39 (.ig/ni3), and S02 (9 ± 4 (.ig/ni3) concentrations.
By using six windows of exposure that covered one to three weeks of gestation, which is all of gestation
in a mouse, these authors (Silva et al., 2008) determined that a significant decrease in near-term fetal
weight (GDI 9) could be induced by ambient air-exposure at least during the first week of gestation.
Decreased placental weight could be induced by ambient air exposure during any of the three weeks of
gestation. These studies point to possible windows of exposure that may be important in evaluating
epidemiologic study results.
Neurodevelopmental Effects
The diagnosis of autism is on the rise in the Western world with its etiology mostly unknown.
Autism is associated cell loss in specific brain regions that is hypothesized to be developmental in origin.
December 2008
7-80
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Sugamata et al. (2006) exposed pregnant ICR mice to DE (0.3 mg DEP/m3) continuously from 2 days
post-coitus (dPC) to 16 dPC. Pups with in utero exposure to DE were nursed in clean air chambers but
may have received gastro-intestinal exposure via lactational transfer of various components of DE. At 11
weeks of age, cerebullar brain tissue was collected. Twenty animals were in each group (10 females, 10
males) with one group receiving clean air exposure and one receiving DE; no filtration was used to
compare PM v. gaseous DE exposure. Earlier work has shown that DEP (<0.1 |im) have been detected in
the brains (cerebral cortex and hippocampus) of newborn pups who were born to dams who were exposed
to DE during pregnancy (Sugamata et al., 2006). Histological analysis of DE-exposed pup cerebella
revealed significant increases in caspase-3 (c-3) positive cells compared to control and significant
decreases in cerebella Purkinje cell numbers in DE-exposed animals versus control. The ratio of cells
positive for apoptosis (c-3 positive) showed a nearly significant sex difference with males displaying
increased apoptosis versus females (p = 0.09). In humans with autism, the cerebellum has a decreased
number of Purkinje cells, which is thought to be fetal and developmental in origin; further, these authors
speculate that humans may be more sensitive to DE-dependent neuronal brain changes as the human
placenta is 2 layers thick whereas the mouse placenta is 4 layers thick.
Behavioral Effects
Body weight decrements at birth have recently been associated through the Barker hypothesis with
adverse adult outcomes. Thus, many publications have begun to focus on decreased birth weight for
gestational age and associated adult changes. Hougaard et al. (2008) exposed 40 timed-pregnant C57BL/6
dams to DEP reference materials (aged DE particulate extract) via inhalation chamber over GD7-19 of
pregnancy. They found significantly decreased pup weight at weaning, albeit not at birth. PM-dependent
liver changes were monitored by following various inflammatory and gentoxicity-related mRNA
transcripts; there were no significant differences in pups at PND2. The comet assay from PND2 pup livers
showed no significant differences between DEP-exposed and control animals. Thyroxine was unchanged
in control and DEP-exposed dams and offspring at weaning. At two months, female DEP-exposed pups
required less time than controls to locate the platform in its new location during the first trial of the spatial
reversal learning task in the Morris water maze (p<0.05). DEP extract exposure during in utero
development led to decreased body weight at weaning and no changes in inflammatory markers, or
thyroid hormone levels.
Lactation
Lactational exposure to various environmental compounds is an area of research that is often
overlooked. Breast milk is a complex matrix, which is essential for the survival of many species.
December 2008
7-81
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Compounds that are especially lipophilic are commonly found in breast milk of exposed dams and this
maternal load can be transferred to the developing neonate. PM in DE adsorbs many chemicals, including
polycyclic aromatic hydrocarbons (PAHs), which have been shown to have to be mutagenic and to be
estrogenic/antiestrogenic and antiandrogenic (Hirose et al., 2001; Kizu et al., 2003). Thus, Tozuka et al.
(2004) monitored the transfer of aromatic hydrocarbons to fetuses and breast milk of Fisher 344 rats
exposed to DE for two weeks from GD7 to GD 20 (minus four days of the weekend with no exposures)
for 6h/day with PMi0 concentration of 1.73 mg/m3. Concentrations of individual PAHs were monitored in
the inhalation chambers including Ace, Fie, Phe, Ant, Flu, Pyr, BaA, Chr, BbF, BkF, BaP, DBA, BghiP,
and IDP at 150 ± 34, 3160 ± 401, 2280 ± 291, 70.3 ± 10.9, 148 ± 19, 133 ± 5, 17.2 ± 2.7, 39.9 ± 6.8,
9.9 ± 2.1, 4.9 ±1.1, 3.7 ± 0.5, <1.4, <6.0, and 4.2 ±0.1 ng/m3, respectively. At PND 14, milk was
collected from exposed and control rats. Fifteen PAHs were monitored in DE-generated air. Seven of
these were quantified in dam blood with levels of phenanthrene (Phe), anthracene (Ant) and
benz[a]anthracene (BaA) in the DE group being significantly higher than control group. In breast milk,
acenaphthene (Ace), fluorene (Fie), Phe, Ant, fluoranthene (Flu), pyrene (Pyr), BaA and chrysene (Chr)
were quantified. Ant, Flu, Pyr and Chr showed significant increases in the DE group compared to control
milk. BaA tended to be about four fold higher than the control group in breast milk, but the increase was
not significant. PAHs in dam livers of DE versus control were not significantly different. PAHs are
transferred across the placenta from the DE-exposed dam to the fetus. Lactational transfer through the
breast milk is also likely as PAHs are detected in dam breast milk, but this should be confirmed in future
studies that cross foster control and exposed dams and pups. The lipophilicity of the PAH based on its
structure affected its uptake to the dam from the air as PAHs with 3 or 4 rings were found in maternal
blood and PAHs with 5 or 6 rings were not detected in dam blood.
7.4.3. Summary and Causal Determination
7.4.3.1. PM10
In summary, epidemiologic studies do not consistently report associations between PMi0 exposure
and preterm birth, growth restriction, birth defects or decreased sperm quality; however evidence is
accumulating for effects on low birth weight and infant mortality, especially due to respiratory causes
during the post-neonatal period. The most striking results were that three U.S. studies reported 11 gram
decrements in birth weight associated with PMi0 exposure. The consistency of these results strengthens
the interpretation that particle exposure may be causally related to reductions in birth weight. Similarly,
animal evidence supported an association between PMi0 exposure and adverse reproductive and
December 2008
7-82
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
developmental outcomes, but provided little mechanistic information or biologic plausibility for an
association between long-term PM exposure and adverse birth outcomes, including low birth weight, or
infant mortality. New evidence from animal toxicological studies on heritable mutations is promising, and
warrants further investigation. Overall, the epidemiologic and toxicological evidence is suggestive of
a causal relationship between long-term exposures to PM10 and reproductive and developmental
outcomes.
The epidemiologic studies generally show consistency and coherence across the different health
outcomes assessed; though interpretation of the results remains challenging given methodological issues
discussed previously. Three epidemiologic studies in the U.S. (Bell et al., 2004; Chen et al., 2002; Salam
et al., 2005) reported 11 gram decrements in birth weight associated with PMi0 exposure. The
consistency of these results, in different studies by different investigators in different regions of the U.S.,
strengthens the interpretation that particle exposure may be causally related to reductions in birth weight.
In epidemiologic studies, all but one of the studies (Wilhelm and Ritz, 2005) that examined the
relationship between PM10 and preterm birth found at least some positive associations for the exposure
periods analyzed. The effects tended to be the strongest for exposures during the first trimester. Three
epidemiologic studies evaluated the association between PM10 and growth restriction; two found no
association (Hansen et al., 2007a; Salam et al., 2005) and one study found a positive association with
exposures during the second trimester (Mannes et al., 2005). The results of three epidemiologic studies
examining PMi0 and birth defects have produced inconsistent results; two found a positive association
(Gilboa et al., 2005; Kim et al., 2007a) and one found inconclusive results (Ritz et al., 2002). Finally,
epidemiologic studies of infant mortality (<1 year) found positive associations, although not all were
statistically significant. For total neonatal deaths (<1 month), again all studies found positive associations,
not including a study that did not provide numerical results for this analysis, although some associations
were uncertain (Ritz et al., 2006). Evidence is strongest for the link between PMi0 and postneonatal (>1
month to <1 year) respiratory-related mortality.
Toxicological studies of female reproductive outcomes provide some evidence of PM-related
effects. Windows of exposure are important in effects seen in reproductive fecundity of female mice.
Neonates or adult female BALB/C mice were exposed to PM (ambient urban air in Sao Paulo), and then
were bred to monitor pregnancy and birth outcomes. Adult exposed animals experienced no adverse
outcomes but, nenonatally exposed females had decreased fertility, decreased number of live born
animals, and an increased number of implantation failures pointing to neonatal exposure to PM being a
more critical window for effects in pregnancy and birth outcomes than adult exposure.
December 2008
7-83
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
In a toxicological study to assess placental weights and birth outcomes, pregnant Swiss mice were
exposed to ambient air in Sao Paulo over multiple portions of gestation, which is 3 weeks in mice, and
fetuses were collected near term. Dams exposed to ambient air during the first week of gestation had pups
with significantly decreased fetal weight. Exposure to ambient air over any portion of gestation induced
significant decreases in placental weight. Birth weight is affected by exposure early in pregnancy whereas
exposure during any portion of pregnancy affects placental weight, pointing to windows of exposure to
consider for birth outcomes.
7.4.3.2. PM2.5
In summary, epidemiologic studies do not consistently report associations between PM exposure
and preterm birth, growth restriction, birth defects or decreased sperm quality; however evidence is
accumulating for effects on low birth weight and infant mortality, especially due to respiratory causes
during the post-neonatal period. Exposure to PM2 5 was usually associated with greater reductions in birth
weight than exposure to PMi0. Similarly, animal evidence supported an association between PM25
exposure and adverse reproductive and developmental outcomes, but provided little mechanistic
information or biologic plausibility for an association between long-term PM exposure and adverse birth
outcomes, including low birth weight, or infant mortality. Overall, the epidemiologic and toxicological
evidence is suggestive of a causal relationship between long-term exposures to PM2.5 and
reproductive and developmental outcomes.
Three epidemiologic studies in the U.S. and one in Europe were able to examine the effects of
PM2 5, and all found an increased risk of low birth weight specifically related to PM2 5 exposure (Bell et
al., 2007b; Parker et al., 2005). Exposure to PM2 5 was usually associated with greater reductions in birth
weight than exposure to PM10. All of the studies that examined the relationship between PM2 5 and
preterm birth report positive associations, though the results from Sagiv et al. (2005) were not
statistically significant. Additionally, Ritz et al. (2007) only found significant positive associations among
a subset of the cohort they were examining, and not for the full cohort. Three studies evaluated the
association between PM2 5 and growth restriction and all three found positive associations, with the
strongest evidence coming when exposure was assessed during the first or second trimester (Liu et al.,
2007d; Mannes et al., 2005; Parker et al., 2005). For infant mortality (<1 year), one study examined PM2 5
and found a positive and statistically significant association (Loomis et al., 1999). Two more recent
studies also found positive, though not statistically significant results (Woodruff et al., 2006; 2008).
In toxicological studies of female reproductive outcomes, adult female Sic: ICR mice exposed to
DEP developed significantly decreased uterine weight versus control. A subset of these adult-exposed
December 2008
7-84
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
animals was bred to unexposed males and the exposed females showed maternal behavioral changes
(decreased rate of good nest construction). The offspring had decrements in body weight, with decreases
in organ weights (females), decreased anogenital distance in males (anti-androgen indicator), and
decreased crown to rump length (females). Adult exposure prior to mating affects offspring development
and maternal behavior. In utero DEP exposure in Sic: ICR mice led to decreased mRNA expression of
BMP-15, a gene related to oocyte development, in female offspring with no changes in the message level
of MIS and SF-1, two sexual differentiation genes that are affected in male offspring with in utero DEP
exposure.
Several studies also provided some limited evidence of male reproductive effects. In utero DEP-
exposure induced changes in male fetal mRNA for MIS and SF-1, two genes important in male sexual
differentiation. These changes were differentially regulated based on the mouse strain used (most
sensitive, ICR > C57BL/6 > ddY, least sensitive).
DE-exposure in utero led to conflicting reports of reproductive outcomes in male offspring (ICR
mice). Some have reported decreased body weight and decrements in spermatogenesis (PND 8-84 at
various time points) including decreased serum testosterone, decreased male accessory reproductive gland
weight, histological vacuolization of Sertoli cells in the seminiferous tubules, and decreased daily sperm
counts. Others showed DE-exposure in utero induced increased serum testosterone and increased
reproductive gland weight and increased body weight in male offspring (ICR mice) at one time period
early in life (PND28) or dose-dependent increased reproductive gland weight, increased serum
testosterone, and decreased expression of mRNA transcripts related to sexual differentiation at 4 weeks of
age. In utero exposure of mice to DE, filtered air (PM removed), or clean air demonstrated that the
gaseous phase of DE was responsible for the decrements in sperm production seen in mice exposed to DE
during gestation.
DE exposure to adult male mice (ICR strain) daily for 6 months led to dose-dependent significant
increases in degradation of seminiferous tubules, and significant decreases in daily sperm production at 6
months of age.
Adult Wistar rats exposed to motorcycle exhaust (MEP) daily for four weeks developed significant
decreases in body weight, spermatids (testis and caudal epididymal), testicular weight, superoxide
dismutase, and serum testosterone with increased expression of inflammatory cytokines. Vitamin E co-
treatment with MEP induced a partial rescue of serum T and sperm counts with a return of inflammatory
cytokines to baseline levels. MEP exposure affected sperm production but this decrement was partially
restored by treatment with the antioxidant Vitamin E.
Toxicological studies also reported effects for several development outcomes:
December 2008
7-85
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
¦ Immunological effects - placenta: Placental insufficiency after in utero DE-exposure was
followed with an emphasis on immunological parameters. Exposed dams developed increased
levels of inflammatory placental cytokines, with dose discrepancies in the number of
placental absorptions. Greater n should be used in future studies to address the effect of DE
exposure on placental absorptions.
¦ Immunological effects related to asthma: Gestational exposure to PM (ROFA leachate by
inhalation or DEP by intratracheal instillation) made offspring (Sic: CR mice) more
susceptible to asthma. There were significant increases in airway hyper-reactivity seen after
aerosol challenge in ROFA or DEP exposed animals. Pregnant mice were also sensitive to
DEP. Pregnancy and in utero exposure to PM enhances future responses to PM in both dams
and offspring.
¦ Neurodevelopmental effecs: Exposure to DE continuously during gestation manifests as
changes in the rodent (ICR mice) cerebellum that resemble changes seen in humans with
autism, i.e., significant decreases in Purkinje cell numbers.
¦ Behavioral effects: Female offspring of dams exposed during gestation to DEP required less
time that control animals to locate platforms in the Morris water maze in spatial reversal
learning task. After in utero DE exposure, thyroid hormone levels were unchanged as were
inflammatory markers in the liver.
7.4.3.3. PMi0-2.5
Evidence is inadequate to determine if a causal relationship exists between long-term
exposure to PM10-2.5 or other PM components and developmental and reproductive outcomes
because studies have not been conducted in sufficient quantity or quality to draw any conclusion.
7.5. Cancer Incidence, Mutagenicity, and Genotoxicity
7.5.1. Epidemiologic Studies
A limited number of epidemiologic studies have evaluated associations between long-term
exposure to PM and incidence of cancer; studies of mortality from cancer are discussed in the following
section. A summary of the mean PM concentrations reported for the studies characterized in this section is
presented in Table 7.6.
December 2008
7-86
DRAFT-DO NOT QUOTE OR CITE
-------
Table 7-6.
Characterization of ambient PM concentrations from select studies of cancer and
long-term exposures.
Reference
Location
Pollutant
Mean Annual
Concentration (|jg/m3>
Upper Percentile
Concentrations (|jg/m3>
Beelen et al. (2008b)
The Netherlands
PM2.5
28.3
Max: 36.8
Bonner et al. (2005)
Western NY State
TSP
44
Pallietal. (2008)
Florence, Italy
PM10
NR
Pedersen et al. (2006) Czech Republic
PM2.5
Max: 46-120
PM10
Max: 120-238.6
Sorensen et al. (2005) Copenhagen, Germany
PM2.5
12.6-20.7
75th: 24.3-27.7
Vinzents et al. (2005)
Copenhagen, Germany
PM10
16.9-23.5
1 Beelen et al. (2008b) looked at the association of BS, PM2 5, and traffic intensity variables
2 (intensity on nearest road, intensity in a 100 meter buffer zone, and an indicator variable for living close
3 to a major road) with lung cancer incidence in the Netherlands Cohort Study of Diet and Cancer
4 (n = 114,378). This portion of the study was conducted between September 1986 and December 1997.
5 Adjusted analyses included 1,940 cases in the full cohort, and 1,295 cases in a case-cohort analysis of the
6 same data source. The results of this study are presented in Table 7-7.
Table 7-7.
Association of average air pollution concentrations and traffic variables with lung
cancer incidence in full cohort and case-cohort analyses.
Pollutant
Increment
Full Cohort
Case-Cohort
Never Smokers
Ex Smokers
Current
Smokers
BS
10 (jg/m3
0.96 (0.83-1.11)
1.03 (0.78-1.34)
1.47(1.01-2.16)
0.91 (0.68-1.23)
0.85 (0.70-1.03)
PM2.5
10 |jm/m3
0.81 (0.63-1.04)
0.65 (0.41-1.04)
Traffic intensity on
nearest road
10,000 rrwh/24 h
1.05 (0.92-1.19)
1.07 (0.84-1.36)
1.11 (0.88-1.41)
0.98 (0.77-1.25)
1.04 (0.91-1.19)
Traffic intensity in a
100-m buffer
335,000 rrwh/24 h
1.05 (0.92-1.19)
1.07 (0.84-1.36)
1.36 (0.99-1.87)
1.06 (0.82-1.38)
0.96 (0.81-1.14)
Living near a major
Road
1.11 (0.91-1.34)
1.10(0.74-1.62)
1.55 (0.98-2.43)
1.24(0.85-1.81)
0.95 (0.73-1.23)
Source: Beelen et al. (2008b)
7 Bonner et al. (2005) conducted a population-based, case-control study of the association between
8 ambient exposure to polycyclic aromatic hydrocarbons (PAHs) in early life and breast cancer incidence
9 among women living in Erie and Niagara counties in the state of New York. Cases (n = 1166) were
10 women with primary breast cancer, and controls (n = 2105) were frequency matched to the cases by age,
11 race, and county of residence. TSP was used as a proxy for PAH exposure. Annual average TSP
December 2008
7-87
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
concentrations (1959-1997) were obtained from the New York State Department of Environmental
Conservation for Erie and Niagara Counties. Among postmenopausal women, exposure to high
concentrations of TSP (>140 |_ig/m3) at birth was associated with an OR of 2.42 (95% CI: 0.97-6.09)
relative to low concentrations of TSP (<84 |_ig/m3). ORs also were elevated for pollution exposures at age
of menarche (OR: 1.45 [95% CI: 0.74-2.87]) and age at first birth (OR: 1.33 [95% CI: 0.87-2.06]) among
postmenopausal women. Among premenopausal women, exposure to high concentrations of TSP at birth
was associated with an OR of 1.79 (95% CI: 0.62-5.10) relative to low exposure levels, exposure at age of
menarche was associated with an OR of 0.66 (95% CI: 0.38-1.16), and exposure at age of first birth was
associated with an OR of 0.52 (95% CI: 0.22-1.20).
Markers of Exposure or Susceptibility
Several researchers conducted studies that looked at markers of exposure or susceptibility as the
outcome associated with short-term exposure. These studies are included here because they may be
relevant to the mechanism that leads to cancer associated with long-term exposures. Investigators
conducted a family pilot study in the Czech Republic to test the hypothesis that exposure to air pollution
with PM in children results in detectable effects indicated by a number of biomarkers of exposure and
early effects (Pedersen et al., 2006). In this pilot study 24 families, from Prachatice and Teplice, were
identified and the mothers and any children age 5-11 were asked to participate. Exposures were assessed
using air sampling at the front door of participants' homes, monitoring data from two stationary air
monitoring stations, biomarkers of exposure, and questionnaires. Repeated short-term air sampling was
performed during five days at participants' residences. Time series of the levels of common air pollutants
were available for each area. Cotinine, a biomarker of nicotine, was measured in samples of urine to
assess exposures to environmental tobacco smoke. Participants also provided 5 ml of blood sample and
the frequency of micronuclei in peripheral blood lymphocytes was analyzed the cytogenetic effects in
study subjects. Significantly higher frequencies of micronuclei were found in younger children living in
Teplice compared to those living in Prachatice. This finding is noteworthy considering micronuclei
formation in peripheral blood lymphocytes is assumed to be biologically relevant for carcinogenesis.
Palli et al. (2008) investigated the correlation between ambient PMi0 concentrations and individual
levels of DNA bulky adducts. Study participants were 214 healthy adults aged 35-64 years at enrollment
who resided in the city of Florence, Italy. This study was conducted between 1993 and 1998. PMi0
exposure levels were based on daily environmental measures provided by two types of urban monitoring
stations (high-traffic and low-traffic). The researchers assessed correlation between DNA bulky adducts
measured in blood samples and PMi0 concentrations prior to blood sample collection. Time windows of
PMio exposure evaluated in this study were 0-5 days, 0-10 days, 0-15 days, 0-30 days, 0-60 days, and
December 2008
7-88
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
0-90 days prior to blood sample collection. Overall, average PM10 concentrations decreased during the
study period, with some fluctuations. Exact values were not reported, but PM10 appeared to range between
approximately 30 and 100 (ig/m3 for high-traffic stations, and between approximately 20 and 50 (ig/m3 for
low-traffic stations. This study found that levels of DNA bulky adducts among non-smoking workers with
occupational traffic exposure were significantly correlated with cumulative PMi0 levels from high-traffic
stations during approximately 2 weeks preceding blood sample collection (0-5 days: r = 0.55, p = 0.03;
0-10 days: r = 0.58, p = 0.02; 0-15 days: r = 0.56, p = 0.02). DNA bulky adducts were not associated
with PMio levels among Florence residents with no occupational exposure to vehicle emissions or among
smokers. DNA bulky adducts were not associated with PMi0 levels assessed by low-traffic urban
monitoring stations.
Sorensen et al. (2005) investigated the association between personal exposure to water-soluble
transition metals in PM2 5 and oxidative stress-induced DNA damage. This study was conducted among
49 students from Central Copenhagen, Denmark. Researchers assessed PM2 5 exposure by personal
sampling over two week-day periods twice in one year (November, 1999 and August, 2000), and
determined the concentration of water-soluble transition metals (vanadium, chromium, iron, nickel,
copper, and platinum) in these samples. In addition, students donated lymphocyte and 24-h urine samples
which were analyzed for DNA damage in terms of 7-hydro-8-oxo-2'-deoxyguanosine (8-oxodG). Mean
concentrations and corresponding IQR of these metals differed between months of sample collection. This
study found that 8-oxodG concentration in lymphocytes was significantly associated with vanadium and
chromium concentrations, with a 1.9% increase in 8-oxodG per 1 |_ig/L increase in vanadium
concentration and a 2.2% increase in 8-oxodG per 1 (ig/L increase in chromium concentration. Vanadium
and chromium were associated with the 8-oxodG concentration in lymphocytes independent of the PM2 5
mass concentration. Platinum, nickel, copper, and iron were not significantly associated with the 8-oxodG
concentration in lymphocytes, and none of the six measured transition metals was associated with the 8-
oxodG concentration in urine.
Vinzents et al. (2005) investigated the association between UFPand PMi0 concentrations with levels
of purine oxidation and strand breaks in DNA using a crossover design. Study participants were 15
healthy nonsmoking individuals with a mean age of 25. UFP exposure was evaluated in terms of number
concentrations in the breathing zone by using portable instruments in six 18-h weekday periods from
March to June 2003. Ambient concentrations for PMi0 and UFP were also measured on all exposure days
at curbside street stations and at one urban background station. Oxidative DNA damage was assessed by
evaluating strand breaks and oxidized purines in mononuclear cells isolated from venous blood the
morning after exposure measurement. Mean concentration of UFPs (street station) was 30.4 x 103
UFPs/mL (standard deviation [SD]: 1.38), mean concentration of PMi0 at a background monitoring station
December 2008
7-89
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
was 16.9 (ig/m3 (SD: 1.53), and mean concentration of PM10 at a street station was 23.5 (ig/m3 (SD: 1.48).
Mean personal exposure to UFPs was 32.4 x 103 UFPs/mL (SD: 1.49) while bicycling (5 occasions), 19.6
x 103 UFPs/mL (SD: 1.78) during other outdoor activities (6 occasions), and 13.4 x 103 UFPs/mL (SD:
1.96) while indoors (6 occasions). The regression coefficients of the mixed-effects models looking at level
of purine oxidation were estimated as 1.50 x 10 (95% CI: 0.59 x 10 to 2.42 x 10 3: p = 0.002) for
cumulated outdoor exposure and 1.07 x 10~3 (95% CI: 0.37 x 10~3 to 1.77 x 10~3; p = 0.003) for
cumulated indoor exposure. Cumulated outdoor and cumulated indoor exposures to UFPs were not
associated with strand breaks. Neither ambient air concentrations of PMi0 nor number concentrations of
UFPs at monitoring stations were significant predictors of DNA damage.
Summary
Though several studies have reported an association between lung cancer mortality and long-term
PM exposure, the single study (2008a) that looked at lung cancer incidence found no association with
PM2 5. There are known constituents of PM that have varying levels of toxicity, including some that have
been classified as possible or probable carcinogens. An epidemiologic study looked at PM constituents
(using TSP as a surrogate for PAHs) and found a positive association. Overall, there is limited evidence
available to evaluate the relationship between relevant PM exposures and cancer incidence, though future
studies of the effects of PM on DNA damage and other precursors to carcinogenesis are warranted.
7.5.1.1. Toxicological Studies
Over the past 30 years numerous mutagenicity and genotoxicity studies of ambient PM and their
contributing sources have been done to assess the relative mutagenic/genotoxic potential and health risks
associated with human exposure to PM. The results from many of these studies were previously described
in the 2004 PM AQCD (U.S. EPA, 2004). Most of the data published since then have focused on acute
cardiovascular or respiratory effects associated with short-term exposure to ambient PM, mobile
combustion sources, or selected constituents. Building on results of earlier studies in the 2004 PM AQCD,
data from newly published studies that evaluated the mutagenic and/or genotoxic effects of PM, PM-
constituents, and combustion emission source particles are reviewed. A summary table of the pertinent
studies is provided in Annex D.
Studies previously reviewed in the 2004 PM AQCD (U.S. EPA, 2004) provide compelling evidence
that ambient PM as well as PM from specific combustion sources (e.g., fossil fuels) is mutagenic in vivo
and in vitro. Studies of neat PM, extracts, and specific components identified in different types of PM
have reported positive results in Salmonella typhimurium mutation tests (Ames), and mutations,
micronuclei (MN), sister chromatid exchange (SCE), chromosomal aberrations, and/or DNA damage in
December 2008
7-90
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
various types of human or laboratory animal cells. One caveat to interpretation of these data is that there
is not a simple linear relationship between mutagenic potential and carcinogenic potential in animals or
humans. Usually studies focus on organic fractions of PM extracts for mutagenicity testing using bacteria
and mammalian cell lines.
PM and/or PM extracts from ambient air samples (e.g., southern California), wood smoke, and
coal, diesel, or gasoline combustion have all been reported to induce mutation in S. typhimurium and in
cultured human cells. The effect of seasonal and spatial factors on the activity of ambient PM appears to
be related to the overall stationary vs. mobile contributory sources. A limited number of studies evaluated
the impact of the season on the genotoxic effects of ambient PM. A few studies however have indicated
that greater genotoxic effects were associated with samples collected during the winter months compared
to those collected in the summer (Abou Chakra et al., 2007; Gabelova et al., 2007a; Gabelova et al.,
2007b). Comparatively, Hannigan et al. (1997) indicated that no seasonal variation was observed. Studies
have also shown that greater genotoxic effects were associated with smaller particle size extracts (e.g.,
PM2 5>PMio) and from samples collected in urban areas or closer to higher trafficked areas (Abou Chakra
et al., 2007; Avogbe et al., 2005; Hornberg et al., 1998).
Studies also have found that unsubstituted polyaromatic compounds were responsible for much of
the mutagenic activity of PM. These compounds included benzo[k]fluoranthene, indeno[l,2,3-cd]pyrene,
benzo[b]fluoranthene, benzo[g,h,i]perylene, benzo[a]pyrene, and cyclopenta[cd]pyrene. Studies of PM
samples from several European locations also reported PM was mutagenic in numerous test systems.
PM and related constituents induced mutations at the HPRT locus in Chinese hamster ovary (CHO)
cells and alveolar type II cells, APRT and Ouar loci in CHO cells, thymidine kinase (TK) locus in the
TK6 human lymphoblast cells and specific-locus mutations in mice (source). DNA damage, unscheduled
DNA synthesis, and SCE were induced in mammalian (e.g., Chinese hamster) cells in vitro (source).
Mutagenicity studies of PM and PM extracts from gasoline- powered engines also reported positive
results for most of these same endpoints, however the activity was generally lower (source).
Test results from studies of point source PM (e.g., wood or coal smoke) often show that organic
extracts of the PM samples have greater mutagenic potency than untreated PM. Similarly, the PM sample
source generally affects mutagenic potency. For example, PM from wood smoke was weakly mutagenic
compared to coal, which was slightly less mutagenic than DEP. Direct-acting frameshift mutations were
induced in S. typhimurium in studies of fluidized-bed combustion fly ash but coal combustion fly ash was
not mutagenic. PM samples and extracts collected from homes in Yunnan Province, China that burn
smoky-coal for heating and cooking were highly mutagenic in bacteria in the presence of S-9 (Mumford
et al., 1987).
December 2008
7-91
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Carcinogenesis
HWS-dependent lung cancer induction was studied in a mouse model that spontaneously develops
lung tumors, strain A/J mice (Reed et al., 2006). These mice were exposed to HWS for 6 months followed
by 6 months of recovery with no HWS exposure. At 14 months, these mice were collected and tumor
multiplicity and tumor incidence were measured. Gaseous components of the HWS included NOx, N02,
CO, S02, NH3, and non-methane volatile organic carbon with concentration from control levels to high
dose HWS exposure ranging from 0 to 0 ppm, 0 to 0 ppm, 229 ± 31 to 14887.6 +/- 832.3 ppm, 0 to 0 ppb,
139.3 +/- 2.3 to 54.9 +/- 1.2 (ig/m3, and 177.6+/-10.4 to 3455.0 +/- 557.2 (.ig/ni3, respectively. Total PM
mass ranged from 6.4+/-6.9 in control groups to 40.5+/-9.2 in the low group to 1041.1+/-123.5 |_ig/m3 in
the high group. Cancer indicators showed no significant differences versus control animals. However,
HWS from this study was mutagenic in the Ames reverse mutation assay. HWS did not enhance lung
cancer development in this rodent model. Similar studies with exposure to environmental levels of DE
revealed similar outcomes, namely no increase in lung adenomas in A/J mice after 6 months of exposure
to DE followed by six months of recovery (Reed et al., 2004). The concentration of gases in this DE
experiment including NOx, N02, CO, S02, NH3, methane, non-methane volatile organic carbon, and FID
total hydrocarbon ranged from control to high dose group values of 0 to 50.4 +/- 0.6 ppm, 0.2 +/- 0.2 to
6.9 +/- 3.3 ppm, 0.3 +/- 0.1 to 30.9 +/- 4.5 ppm, not detectable to 955.2 +/- 58.4 ppb, 176.5 +/- 8.8 to
9.1 +/- 0.2 (ig/m3, 1406.5 +/-253.2 to 2642.1 +/- 455.9 (ig/m3, 134.0 +/- 52.1 to 1578.6 +/- 256.2 (ig/m3,
0.1 +/- 0.1 to 2.2 +/- 0.2 ppm, respectively. Total PM mass was 8.7+/-8.5 |_ig/m3in controls, 43.6+/-
8.4 (ig/m3 in low dose, and 1005.0+/-74.6 (ig/m3 in high dose exposures. Micronucleated reticulocytes
(MN), a genotoxicity marker, did not differ between DE-exposed and control groups. Exposure to
environmentally relevant concentrations of DE or HWS did not cause an increased rate of lung tumors in
a rodent model of lung cancer.
DNA Damage
Sato et al. (2003a) examined DNA adduct formation in lungs, nasal mucosa and livers of adult
male Wistar rats exposed to ambient urban roadside air for 4, 12, 24, 48, or 60 weeks in Kawasaki, Japan
(1995-1996). They also monitored message levels of cytochrome P450 (CYP) enzymes that catalyze the
transformation of PAHs to reactive metabolites. PM was measured as suspended PM (SPM).
Concentrations of gases were reported to be 12-182 ppb NO and 0-9 ppb N02 in the filtered air chamber
and 33-280 ppb NO and 42-81 ppb N02 in the experimental group chamber. SPM concentrations were
reported to be 11-19 |_ig/m3 in the filtered air chamber and 42-100 (ig/m3 (average 63 |_ig/m3) in the
experimental group chamber. Body weight significantly decreased in exposed animals at 24, 48 and 60
weeks. With the most acute exposure of 4 weeks, there were significant increases in multiple DNA
December 2008
7-92
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
adducts (lung, nasal, and liver DNA adducts). With longer exposures, there were significant increases in
lung (48 weeks), nasal (60 weeks), and liver DNA adducts (60 weeks). Changes were seen in CYP1A2
mRNA at 4 weeks with 2.3 fold increased message level in exposed animals compared to the control
group with no change seen at 60 weeks; at 4 and 60 weeks, CYP1A1 was unchanged. These results
indicate that exposure to ambient air in this roadside area can induce DNA adduct formation, which may
be important for carcinogenicity as earlier studies (Ichinose et al., 1997) have shown that 8-oxo-dG is
elevated along with tumor formation in a dose-dependent manner in mice administered diesel particles.
The finding of adducts in the liver indicate that deposition of PM and its associated PAHs in the lung can
have effects at downstream organ (liver). However, PM deposition on the fur and ingestion during
grooming cannot be ruled out as a possible exposure route.
Mutagenesis and Genotoxicity
The specific effects produced by PM and particle constituents include induction of MN formation,
DNA adduct formation, SCE, DNA strand breaks, frameshifts and inhibition of gap-junctional
intercellular communication (Alink et al., 1998; Arlt et al., 2007; Avogbe et al., 2005; Gabelova et al.,
2007a; Gabelova et al., 2007b; Healey et al., 2006; Hornberg and Seemayer, 1996; Hornberg et al., 1998;
Sevastyanova et al., 2007).
Assessment of the constituents adsorbed onto individual particles played a significant role in the
genotoxic potential of PM. Studies by Poma et al. (2006) showed that the genotoxic potential of fine
carbon black particles was consistently less genotoxic than similar concentrations of PM2 5 extracts,
suggesting that the adsorbed components play a role in the genotoxic potential of PM. Studies indicated
that total PAH and carcinogenic PAH content is correlated with the genotoxic effects of PM (de Kok et
al., 2005; Sevastyanova et al., 2007). Comparison of different extracts (water vs. organic) by Gutierrez-
Castillo et al. (2006) indicated that water soluble extracts were more genotoxic than the corresponding
organic extracts. Sharma et al. (2007) reported that mutagenic activity of samples collected in and around
a waste incineration plant was found mostly in the moderately polar and polar fractions of filter extracts.
No mutagenic activity was observed from any of the nonpolar samples evaluated. The polar and crude
fractions were mutagenic without metabolic activation, suggesting a direct mutagenic effect. Arlt and
colleagues (2007) have shown that the known PM constituents 2-nitrobenzanthrone and 3-
nitrobenzanthrone were genotoxic in a variety of bacterial and mammalian cell systems.
Conflicting data have been reported for the role of metabolic enzymes on the genotoxicity of PM
and their adsorbed constituents. Arlt et al. (2007) reported that the PM constituent 2-nitrobenzanthrone (2-
NB) was genotoxic in bacterial and mammalian cells. However, metabolic activation with the human N-
acetyltransferase 2 or SULT1A1 enzyme was needed for the effect to be observed in human cells.
December 2008
7-93
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Erdinger et al. (2005) demonstrated that mutagenic activity was not affected when metabolism was
induced, de Kok et al. (2005) evaluated the relationship between the physical, chemical, and genotoxic
effects of ambient PM. TSP, PM10, and PM2 5 were sampled at different locations and the organic extracts
were assessed for mutagenicity and induction of DNA adducts in cells. Overall, induction of rat liver S9
metabolism generally reduced the mutagenic potential via the Ames assay of the particle fractions and
DNA reactivity (induction of DNA adducts) was generally higher after metabolic activation. Binkova et
al. (2003) found that the addition of S9 increased PMio-dependent DNA adduct formation.
Bacterial Test Systems
Wood smoke
The mutagenicity of wood smoke (WS) and cigarette smoke (CS) extracts was assayed in
Salmonella typhimurium strains TA98 and TA100 (Ames assay) using the pre-incubation assay with
exogenous metabolic activation (rat liver S-9). Extracts of both samples (62.5 or 125 |_ig TPM
equivalent/ml) were equally mutagenic to strain TA98 but the WS extract was less mutagenic than the CS
extracts in strain TA100 (Iba et al., 2006).
Traffic-related Ambient Air
de Kok et al. (de Kok et al., 2005) found the direct mutagenicity (Ames assay) and the direct
DNA reactivity (DNA adduct formation) of the PM2 5 size fraction was significantly higher than that of the
larger size fractions (TSP, PM10) at most locations.
Diesel and Gasoline
Automobile DEP (A-DEP) was tested in S. typhimurium strains TA98, TA100, and its derivatives
(e.g., TA98NR and YG1021) and found to be more mutagenic than forklift DEP (i.e., SRM2975)
particles, based on PM mass. A-DEP had 227 times more PAH-type mutagenic activity and 8-45 times
more nitroarene-type mutagenic activity due to the different conditions for generating and collecting the
two DEP samples (DeMarini et al., 2004). Using a diesel engine without an oxidation catalytic converter
(OCC), the diesel engine exhaust particle (DEP) extract produced the highest number of revertant
colonies in strains TA98 and TA100 with and without S9 at several tested loads when compared to
extracts from low-sulfur diesel fuel (LSDF), rapeseed oil methyl ester (RME), and soybean oil methyl
ester (SME). When an OCC was installed in the exhaust pipe of the engine, all extracts reduced the
number of revertant colonies in both strains with and without S9 at partial loads but increased the number
of revertant colonies without S9 at rated power. At idling, DEP extracts increased the number of revertant
December 2008
7-94
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
colonies with and without S9 (Bunger et al., 2006). In a separate study, engine emissions (particle extracts
and condensates) from rapeseed (canola) oil were found to produce greater mutagenic effects in S.
typhimurium strains TA98 and TA100 than DEP (Bunger et al., 2007a). Additionally, DE extract (DEE)
from diesel fuel containing various percentages of ethanol was also observed to induce mutational
response in two Salmonella strains. Base diesel fuel DEE and DEE from fuel with 20% ethanol caused
more significant DNA damage in rat fibrocytes L-929 cells than extracts containing 5, 10, or 15% ethanol
(Song et al., 2007).
DE and gasoline engine exhaust particles, as well as their SVOC extracts, induced mutations in the
two S. typhimurium strains YG1024 and YG1029 in the absence and presence of S9; the PM extracts were
more mutagenic than the SVOC extracts. Additionally, all extracts except the DE SVOC extract induced
DNA damage and MN in Chinese hamster lung V79 cells (Liu et al., 2005b).
In an early study, Lofroth (1981) compared the mutagenic activity of PM from diesel and gasoline
engine exhaust and found both to be mutagenic in the Ames assay, in the absence of mammalian
metabolic activation; the DE mutagenic response was far greater than the mutagenic response of gasoline.
Another older study demonstrated low mutagenicity in cars burning liquified petroleum and cars with
catalysts and high mutagenicity (in TA100 ± S9) in light-duty diesel vehicles (Rannug et al., 1983). Also,
more mutagenesis was observed in exhaust from cold starts (0 °C) than in starts at 23 °C. Another study
demonstrated that gasoline engine exhaust significantly increased colony formation in TA98 with and
without S9 (Zhang et al., 2007b).
Strandell et al. (1994) fractionated the extracts of gasoline and DE from Volvos to find the most
potent mutagens among the subtractions. Mutagenicity testing was done with the Ames assay with strain
TA98, both with and without S9 metabolic activation, and with strain TA98NR (a nitro reductase-
deficient strain used to determine the presence of nitro aromatic mutagens). The most polar subtraction
was also the most mutagenic and comprised 51% of the total mutagenicity for gasoline exhaust and 39%
of the total for DE. This fraction contained low-boiling point components and some phenol derivatives.
Both fuels had a similar TA98NR ± S9 response that was less than TA98, but differed significantly in
their TA98 ± S9 response, which suggests difference in some nitro-reductase-dependent mutagens. A
reduction in mutagenicity was observed with the addition of S9 activation, which the authors attributed to
possible enzymatic deactivation of direct-acting mutagens or activation of uncharacterized compounds.
In Vitro Test Systems
Ambient PM
DNA damage was assessed by the Comet assay in A549 cells exposed to PM collected from a high
traffic area in Copenhagen, Denmark (TSP approximately 30 |_ig/m3) and compared to the results from
December 2008
7-95
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
exposure of A549 cells to standard reference materials (SRM1650 or SRM2975) at the same
concentrations (2.5-250 (ig/ml) (Danielsen et al., 2008a). All three particles induced strand breaks and
oxidized purines in a dose-dependent manner and there were no obvious differences in potency. In
contrast, only the ambient PM formed 8-oxo-7,8-dihydro-2'-deoxyguanosine (8-oxodG) when incubated
with calf thymus DNA and the authors suggested that this may be due to the levels of transition metals.
Woodsmoke
One recent study measured the effect of freshly generated red oak WS on CYP1A1 activity based
on ethoxyresorufin O-deethylase in pulmonary microsomes recovered from male Sprague-Dawley rats
exposed to WS by nose-only inhalation exposure (Iba et al., 2006). CYP1A1 activity in rat lung explants
treated with extracts of the total PM (TPM) from WS samples and from freshly generated cigarette smoke
(CS) was also evaluated. Unlike CS, WS did not induce pulmonary CYP1A1 activity or mRNA (assessed
by northern blot analysis) nor did extracts of WS TPM induce CYP1A1 protein (assessed by western blot
analysis) in cultured rat lung explants. The results suggest that unique constituents that are activated by
CYP1A1 may be present in CS but not WS that are mutagenic for both WS and CS in S. typhimurium
strain TA100.
Diesel and Gasoline
Jacobsen et al. (2008) used the FEl-Muta™ Mouse lung epithelial cell line to investigate putative
mechanisms of DEP-induced mutagenicity. Mutation ion frequencies and ROS were determined after
cells were incubated with 37.5 or 75 (ig/ml DEP (SRM1650) for 72-h (n = 8). The mutation frequency at
the 75 (ig/ml dose was significantly increased (1.55-fold; p<0.001) in contrast to cells treated with
37.5 (ig/ml DEP. DEP induced ROS generation 1.6-1.9-fold in the epithelial cell cultures after 3 h of
exposure compared with the 3-10-fold increase in ROS production previously reported for carbon black.
The authors concluded that the mutagenic activity of DEP is likely attributable to activity from the
organic fraction that both contains reactive species and can generate ROS.
In human A549 and CHO-K1 cells, the organic fraction of DEP significantly increased the amount
of Comet and MN formation, respectively, in the presence and absence of SKF-525A (a CYP450
inhibitor) and S9, respectively (Oh and Chung, 2006). The organic base and neutral fractions of DEP also
significantly induced DNA damage but only without SKF-525A, and all fractions but the moderately
polar fraction (phthalates and PAH oxyderivatives) induced MN formation with and without S9 (Bao et
al., 2007). Gasoline engine exhaust significantly induced DNA damage as measured in the Comet assay
and increased the frequency of MN in human A549 cells (Zhang et al., 2007b). In human-hamster hybrid
December 2008
7-96
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
(Al) cells, DEP (SRM 2975) dose-dependently increased the mutation yield at the CD59 locus; this was
significantly reduced by simultaneous treatment with phagocytosis inhibitors (Bao et al., 2007).
In Vivo Test Systems
Ambient PM
After earlier work showed increased germline mutation rates in herring gulls nesting near steel
mills on Lake Ontario (Yauk and Quinn, 1996) further work was conducted to address air-dependent
contribution to germline mutations by housing male and female Swiss Webster mice in the same location
and comparing mutation rates in those animals with mutation rates of animals housed in a rural setting
with less air pollution (Somers et al., 2002). Six to eight week old mice received 10 week exposures to
ambient air with a six week break after exposure before breeding to allow for sperm maturation (i.e. so
sperm used in fertilization would be generated during ambient air exposures); DNA from pups, dams and
males used in breeding was collected at PND5. These studies showed heritable mutation frequency was
significantly increased (1.5- to 2-fold) when compared to the rural site. Further, these studies confirmed
that this increased mutation frequency was primarily due to increases in mutation frequency mediated
through the paternal germline.
One study by Somers et al. (2004) showed that offspring of mice exposed to ambient air in urban
regions inherited paternal origin expanded simple tandem repeat (ESTR) mutations 1.9- to 2.1-times more
frequently than offspring of mice exposed to HEPA filtered air or those in exposed to rural ambient air.
Mouse expanded simple tandem repeat (ESTR) DNA is composed of short base pair repeats which are
unstable in the germline and tend to mutate by insertion or deletion of repeat units. In vivo and in situ
studies have shown that murine ESTR loci are susceptible to ionizing radiation, and other environmental
mutatgen-dependent germline mutations, and are thus good markers of exposure to environmental
contaminants.
To determine if PM or the gaseous phase of the urban air was responsible for these heritable
mutations, Yauk et al. (2008) exposed mature male C57BlxCBA F1 hybrid mice to either HEPA-filtered
air or to ambient air in Hamilton, Ontario, Canada for three, ten, or ten weeks plus 6 weeks of clean air
exposure. Sperm DNA was monitored for expanded simple tandem repeat (ESTR) mutations, testicular
sample bulky DNA adducts, and DNA single or double strand breaks. This area in Hamilton is near 2 steel
mills and a major highway. Air composition provided by the Ontario Ministry of the Environment showed
TSP concentration of 9.38 ± 17 (ig/m3, PAH concentration of 8.3 ±1.7 ng/m3, and metal at
3.6 ± 0.7mg/m3. Mutation frequency at ESTR Ms6-hm locus in sperm DNA from mice exposed 3 or 10
weeks did not show elevated ESTR mutation frequencies, but there was a significant increase in ESTR
mutation frequency at 16 weeks in ambient air exposed males versus HEPA-filter exposed animals,
December 2008
7-97
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
pointing to a PM-dependent mechanism of action. No detectable adducts were observed in testes samples
at any of the time points monitored. To verify inhalation exposure to particles, DNA adducts were
monitored in the lungs of exposed mice and at 3 weeks, ambient-air exposed mice showed significant
increases in lung DNA adducts versus control (filtered-air exposed animals); no other time points showed
detectable DNA adduct formation. Thus, these studies indicate that the ambient PM phase and not the
gaseous phase is responsible for the increased frequency of heritable DNA mutations.
Diesel
An in vivo study employed gtp delta transgenic mice carrying the lambda EG 10 on each
Chromosome 17 from a C57BL/6J background to investigate the effects of DEP on mutation frequency
(Hashimoto et al., 2007). Mice were exposed via inhalation to DEP or via IT instillation to DEP or DEP
extract and lambda EG10 phages were rescued; E. coli YG6020 was infected with the phage and screened
for 6-thioguanine resistance. The mutagenic potency (mutation frequency per mg) caused by DEP extract
was twice that of DEP, suggesting that the mutagenicity of DEP is attributed primarily to compounds in
the extract, since ~50% of the weight of DEP was provided by the extract. Interestingly, there was no
difference in mutation frequency between the 1 and 3 mg/m3 DEP groups after 12 weeks of exposure.
Summary
A number of recent in vivo and in vitro studies indicate that ambient urban PM is mutagenic.
Research evaluating the mutagenicity of ambient PM from the Los Angeles area has pointed to ubiquitous
emission sources as being responsible for mutagenic activity observed in vitro (Hannigan et al., 1997;
1998). Fractionation of these ambient samples and subsequent mutagenicity assessment has indicated that
six unsubstituted polyaromatic compounds and two semi-polar compounds are the likely mutagens.
Mutagenicity of urban air from Germany has also shown (Hornberg and Seemayer, 1996; Hornberg et al.,
1998; Seemayer and Hornberg, 1998), evidence that the fine fraction of PM exerted greater toxicity.
Additionally, ambient PM from high traffic areas in the Netherlands also induced genotoxic activity.
Emissions from wood/biomass burning have been shown to be mutagenic. Characterization of
wood smoke fractions to assign mutagenicity has shown that the organic fraction is mutagenic and that
the condensate is not. Wood smoke emissions can induce both frameshift and base pair mutations but
have not yet been shown to produce DNA adducts.
Emissions from coal combustion have been shown to be mutagenic, especially the polar and
aromatic fractions. Recent work characterizing the mechanism of genotoxicity has examined the mutation
spectra of coal smoke emissions from Chinese homes burning smoky coal (Granville et al., 2003).
Sequencing the revertants has shown that the mutations in Salmonella exposed to coal smoke extract are
December 2008
7-98
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
similar to mutations seen in lung tumors of women exposed environmentally to the coal smoke, which
differs in notable ways from coal combustion emissions in the U.S.
Extensive studies have demonstrated mutagenic activity in both the particle and gaseous fractions
of DE. By sequential fractionation of DE, apportionment of the mutagenicity is possible, which has
implicated nitrated polynuclear aromatic compounds as being responsible for a substantial portion of the
mutagenicity. Other mutagenically active compounds include ethylene, benzene, 1,3-butadiene, acrolein,
and several PAHs in the gas phase. In addition to Ames assay studies, the induction of gene mutations has
been reported in several in vitro mammalian cell lines after exposure to extracts of DPM. Structural
chromosome aberrations and SCE in mammalian cells have been induced by DE particles and extracts.
Early studies comparing the mutagenicity of gasoline and DE showed that the PM component of
the exhaust is more mutagenic than the condensate fraction, and that overall, DE is more mutagenic than
gasoline exhaust. More mutagenicity is also observed in exhaust from cold starts than from exhausts at
room temperature. Examining the fractional mutagenicity of gasoline and DEs, it was shown that, as with
coal smoke, the polar component has the most mutagenicity, and further, that nitro-PAH is present in the
fraction. A comprehensive study comparing gasoline and DE genotoxicity, using both the PM and SVOC
fractions, demonstrated that both exhausts are mutagenic, but, in general, DE is more mutagenic. Further,
the study implicates PAH and nitroarenes in the genotoxicity. Another current study corroborates these
finding, and includes data suggesting that DNA adduct formation is a component of the mutagenicity.
Exact comparisons of the mutagenicity of combustion emissions of these fuels are not possible
because data provided in the studies vary so greatly in units in which mutagenicity is expressed. Thus,
there is qualitative evidence for the mutagenic/genotoxic potential of both ambient PM and some fuel
combustion products. Many of the published in vitro studies failed to provide details regarding the dose of
PM extract delivered to the cells in vitro. In general, equal volumes of air or amounts of time were
sampled and reported, but only limited, if any, characterization of the amount of PM mass or size was
done or reported in many studies. Thus, any quantitative extrapolation of the reported findings would be
quite difficult. Nevertheless, they collectively do appear to provide some evidence tending to substantiate
the biologic plausibility of potential epidemiologic associations between long-term human exposure to
ambient PM and lung cancer at a cellular level.
The potential for the non-organic constituents of particles to induce oxidative DNA damage has
been less well studied but could also provide support for modes of action for mutagenicity or genotoxicity
of long-term PM exposures. A few studies have demonstrated that redox cycling of persistent quinoid
radicals and Fenton reactions by transition metals in PM can generate ROS that induce DNA damage and
are likely to be key events that contribute to the cytotoxic and carcinogenic potential of PM (Valavanidis
et al., 2005).
December 2008
7-99
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
7.5.2. Summary and Causal Determinations
In summary, only one epidemiologic study examined the effect of a specific size fraction of PM
(PM2 5) and cancer incidence (Beelen et al., 2008a), and found no evidence of an association with lung
cancer. An additional study used TSP as a surrogate for PAHs and found an elevated, though not
statistically significant, risk for breast cancer. Similarly, animal toxicological studies did not focus on
specific size fractions of PM, but rather conducted studies of ambient PM, woodsmoke, and DEP. A
number of recent studies indicate that ambient urban PM, emissions from wood/biomass burning,
emissions from coal combustion, and gasoline and DE are mutagenic and that PAHs and nitroarenes are
genotoxic. Due to the nonspecific measure of PM size fractions in epidemiologic and animal toxicological
studies and the limited and inconsistent epidemiologic studies, the evidence is inadequate to determine
if a causal relationship exists between relevant PM10, PM2.5, PM10-2.5, or ultrafine exposures and
incident cases of cancer.
7.6. Mortality Associated with Long-term Exposure
7.6.1. Review of 1996 and 2004 PM AQCDs
In the 1996 PM AQCD, results were presented for three prospective cohort studies of adult
populations: the Six Cities Study (Dockery et al., 1993); the ACS Study (Pope et al., 1995); and, the
California Seventh Day Adventist (AHSMOG) Study (Abbey et al., 1995). The 1996 AQCD concluded
that the chronic exposure studies, taken together, suggested associations between increases in mortality
and long-term exposure to fine PM (U.S. EPA, 1996).
Discussions of mortality and long-term exposure to PM in the 2004 PM AQCD emphasized the
results of four U.S. prospective cohort studies, but the greatest weight was placed on the findings of the
American Cancer Society (ACS) and the Harvard Six Cities studies, which had undergone extensive
independent reanalysis, and which were based on cohorts that were broadly representative of the U.S.
population. The 2004 PM AQCD concluded that the results from the Seventh-Day Adventist (AHSMOG)
cohort provided some suggestive (but less conclusive) evidence for associations, while results from the
Veterans Cohort provided inconsistent evidence for associations between long-term exposures to PM2 5
and mortality. Collectively, the 2004 PM AQCD found that these studies provided strong evidence that
long-term exposure to PM2 5 was associated with increased risk of human mortality. Effect estimates for
all-cause mortality ranged from 6 to 13% increased risk per 10 (ig/m3 PM2 5, while effect estimates for
December 2008
7-100
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
cardiopulmonary mortality range from 6 to 19% per 10 (ig/m3 PM2 5. For lung cancer mortality, the effect
estimate was a 13% increase per 10 (ig/m3 PM2 5, based upon the results of the extended analysis from the
ACS cohort (Pope et al., 2002). With regard to thoracic coarse particles, the 2004 PM AQCD reported
that no association was observed between mortality and long-term exposure to PMi0.2.5 in the ACS study
(Pope et al., 2002), while a positive but statistically non-significant association was reported in males in
the AHSMOG cohort (McDonnell et al., 2000). Thus, the 2004 PM AQCD concluded that there was
insufficient evidence for associations between long-term exposure to thoracic coarse particles and
mortality. Overall, the 2004 PM AQCD concluded that there was strong epidemiologic evidence for
associations between long-term exposures to PM2 5 and/or sulfates and excess all-cause and
cardiopulmonary mortality.
At the time of the 2004 PM AQCD, only a limited number of the chronic-exposure cohort studies
had considered direct measurements of constituents of PM. With regard to source-oriented evaluations of
mortality associations with long-term exposure, the 2004 PM AQCD noted only the study by Hoek et al.
(2002), in which the authors concluded that long-term exposure to traffic-related air pollution may
shorten life expectancy. However, Hoek et al. (2002) also noted that living near a major road might
include other factors that contribute to mortality associations. There was not sufficient evidence at the
time of the 2004 PM AQCD to draw conclusions on effects associated with specific components or
sources of PM.
The following sections will summarize the science since the previous PM AQCD (U.S. EPA,
2004), and build upon the conclusions of that document to reflect more recent key studies for the PM size
components, constituents, and emission sources, as available. New epidemiologic evidence reports a
consistant association between long-term exposure to PM2 5 and increased risk of mortality. There is little
evidence for the long-term effects of PMi0 and PMi0.2.5 on mortality. Although this section focuses on
mortality outcomes in response to long-term exposure to PM, it does not evaluate studies that examine the
association between PM and infant mortality. These studies are evaluated in Section 7.5: "Reproductive,
developmental, prenatal and neonatal outcomes associated with long-term exposure to PM" because it is
possible that in utero exposures contribute to infant mortality. A summary of the mean PM concentrations
reported for the studies characterized in this section is presented in Table 7-8.
December 2008
7-101
DRAFT-DO NOT QUOTE OR CITE
-------
Table 7-8. Characterization of ambient PM concentrations from studies of mortality and long-
term exposures.
Reference
Location
Mean Annual Concentration (|jg/m3>
Upper Percentile Concentrations (|jg/m3>
PMw
Chen et al. (2005a)
Multicity, CA
52.6
Gehring et al. (2006)
North Rhine, Germany
43.7-48.0
Max: 52.5-56.1
Goss et al. (2004)
U.S.
24.8
75th: 28.9
Puett et al. (2008)
NE U.S.
21.6
PMis
Chen et al. (2005a)
Multicity, CA
29.0
Eftim et al. (2008)
U.S.
13.6-14.1
Max: 19.1-25.1
Enstrom 2005 (2005)
CA
23.4
Max: 36.1
Goss et al. (2004)
U.S.
13.7
75th: 15.9
Janes et al. (2007)
U.S.
Jerrett et al. (2005b)
Los Angeles, CA
Laden et al. (2006)
Multicity, U.S.
10.2-29.0
Miller et al. (2007b)
U.S.
13.4
75th: 18.3
Max: 28.3
Pope et al. (2004b)
U.S.
17.1
Schwartz et al. (2008)
Multicity, U.S.
17.5
Max: 40
Zeger et al. (2007)
U.S.
PMlO-25
Chen et al. (2005a)
Multicity, CA
25.4
Lipfert et al. (2006a)
U.S.
15.0
Max: 25.0
7.6.2. PM2.5
1 Studies since the last PM AQCD include results of new analyses and insights for the ACS and
2 Harvard Six Cities studies, further analyses from the AHSMOG and Veterans study cohorts, as well as
3 analyses of a Cystic Fibrosis cohort and a subset of the ACS from California. The historical and more
4 recent results of both the ACS and the Harvard Six Cities studies are compiled in Figure 7-7. Moreover,
5 since the last PM AQCD, there is a major new cohort analyzed in the literature: the Women's Health
6 Initiative (WHI) study (2007b). Most recently, an ecological cohort study of the nation's Medicare
7 population has also been completed (Eftim et al., 2008). These new findings further strengthen the
8 evidence linking long-term exposure to PM2 5 and mortality, while providing indications that the
9 magnitude of the PM2 5 -mortality association is larger than previously estimated (Figure 7-8).
December 2008
7-102
DRAFT-DO NOT QUOTE OR CITE
-------
Cohort
Study
Years
scs
ACS
SCS
ACS
SCS
ACS
ACS
SCS
ACS
SCS
ACS
Original
Reanalysis
Temporal Changes
Extended
6-Cities Medicare
Dockery et al. 1993 1974-1991
Krewski et al. 2000 1974-1991
Villeneuve et al. 2002 1974-1991
Laden etal. 2006 1974-1998
Eftim et al. 2008 2000-2002
Original
Pope etal. 1995
1982-1989
Reanalysis
Krewski et al. 2000
1982-1989
Extended
Pope etal. 2002
1979-1983
Extended
Pope etal. 2002
1999-2000
Intra-metro LA
Jerrett et al. 2005
1982-2000
ACS Medicare
Eftim et al. 2008
2000-2002
Original
Dockery etal. 1993
1974-1991
Reanalysis
Krewski et al. 2000
1974-1991
Original
Pope etal. 1995
1982-1989
Reanalysis
Krewski et al. 2000
1982-1989
Extended
Pope et al. 2002
1979-1983
Extended
Pope et al. 2002
1999-2000
Intra-metro LA
Jerrett et al. 2005
1982-2000
Extended
Laden ei al. 2006
1974-1998
Reanalysis
Krewski et al. 2000
1982-1989
Extended
Pope etal. 2004
1982-2000
Extended
Pope etal. 2004
1982-2000
Intra-metro LA
Jerrett et al. 2005
1982-2000
Original
Dockery et al. 1993
1974-1991
Reanalysis
Krewski et al. 2000
1974-1991
Extended
Laden et al. 2006
1974-1998
Original
Pope etal. 1995
1982-1989
Extended
Pope etal. 2002
1979-1983
Extended
Pope et al. 2002
1999-2000
Intra-metro LA
Jerrett et al. 2005
1982-2000
Extended
Laden et al. 2006
1974-1998
Extended
Pope et al. 2002
1979-1983
Extended
Pope et al. 2002
1999-2000
Intra-metro LA
Jerrett et al. 2005
1982-2000
All Causes
CPD
CVD
IHD
Lung Cancer
0.75
1.25 1.75
Effect Estimate
2.25
Figure 7-7.
Mortality risk estimates associated with long-term exposure to PM2.5 from the Harvard
Six Cities Study (SCS) and the American Cancer Society Study (ACS).
December 2008
7-103
DRAFT-DO NOT QUOTE OR CITE
-------
Study Cohort
Range of Estimates in 2004 AQCD
McDonnell et al. 2000
Laden et al. 2006
Enstrom 2005
Enstrom 2005
Enstrom 2005
Jerrett et al. 2005
Eftitn et al. 2008
Eftim et al. 2008
Lipfert et al 2006a
Lipfert et al. 20061)
Goss et al 2004
Beelen et al 2008
Beelen et al. 2008
Zeger et al. 2007
Pope et al. 2004
Miller et al, 2007
Laden et al. 20O6
Beelen et al 2008
Beelen et al. 2008
Naess et al 2007
Naess et al. 2007
Naess et al 2007
Naess et al 2007
Chen et al. 2005
Chen et al. 2005
Pope et al, 2004
Jerrett et al. 2005
McDonnell et al. 2000
Jerrett et al. 2005
Laden et al. 2006
Beelen et al. 2008
Beelen et al. 2008
Laden et al. 2006
Beelen et al. 2008
Beelen et al. 2008
Jerrett et al. 2005
McDonnell et al 2000
Naess et al. 2007
Naess et al. 2007
Naess et al. 2007
Naess et al 2007
Laden et al. 2006
Beelen et al. 2008
Beelen et al. 2008
Jerrett et al. 2005
AHSMOG
Harvard 6-Cities
CA Cancer Prevention
CA Cancer Prevention
CA Cancer Prevention
ACS-LA
Medicare Cohort
Medicare Cohort
Veterans Cdiort
Veterans Cohort
U.S. Cystic Fibrosis
NLCS-AIR
NLCS-AIR
MCAPS
ACS
WHI
Harvard 6-Cities
NLCS-AIR
NLCS-AIR
Oslo, Norway
Oslo, Norway
Oslo, Norway
Oslo, Norway
AHSMOG
AHSMOG
ACS
ACS-LA
AHSMOG
ACS-LA
Harvard 6-Cities
NLCS-AIR
NLCS-AIR
Harvard 6-Cities
NLCS-AIR
NLCS-AIR
ACS LA
AHSMOG
Oslo, Norway
Oslo, Norway
Oslo, Norway
Oslo, Norway
Harvard 6-Cities
NLCS-AIR
NLCS-AIR
ACS-LA
Years
1973-1977
1974-1998
1973-1962
1983-2002
1973-2002
1982-2000
2000-2002
2000-2002
1989-1996
1997-2001
1999-2000
1987-1996
1987-1996
2000-2002
1982-2000
1994-1998
1974-1998
1987-1996
1987-1996
1992-1998
1992-1998
1992-1998
1992-1998
1973-1998
1973-1998
1982-2000
1982-2000
1973-1977
1982-2000
1974-1998
1987-1996
1987-1996
1974-1998
1987-1996
1987-1996
1982-2000
1973-1977
1992-1998
1992-1998
1992-1998
1992-1998
1974-1998
1987-1996
1987-1996
1982-2000
All Cause
CV
CHD
IHD
CPD
Respiratory
Lung Cancer
Other
0,5
-1 1
1,0 1,5 2.0
Effect Estimate
2.5
Figure 7-8. Mortality risk estimates associated with long-term exposure to PM2.5 in cohort studies.
1 Harvard Six Cities: A follow-up study has used updated air pollution and mortality data; an
2 additional 1,368 deaths occurred during the follow-up period (1990-1998) vs. 1,364 deaths in the original
3 study period (1974-1989) (Laden et al., 2006). Statistically significant associations are reported between
4 long-term exposure to PM2 5 and mortality for data for the two periods (RR =1.16 [95% CI: 1.07-1.26]
5 per 10 (.tg/m" PM2 5). Of note, however, is a statistically significant reduction in mortality risk reported
6 with reduced long-term fine particle concentrations (RR= 0.73 [95% CI: 0.57-0.95] per 10 ug/m' PM25)
7 This is equivalent to an RR of 1.27 for reduced mortality risks. This reduced mortality risk was observed
December 2008
7-104
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
for deaths due to cardiovascular and respiratory causes, but not for lung cancer deaths. The PM2 5
concentrations for recent years were estimated from visibility data, which introduces some uncertainty in
the interpretation of the results from this study. Coupled with the results of the original analysis (Dockery
et al., 1993), this study strongly suggests that a reduction in fine PM pollution yields positive health
benefits.
ACS Extended Analyses: One new analysis further evaluated the associations of long-term PM25
and sulfate exposures with risk of mortality in 50 U.S. cities reported by Pope and colleagues (2002),
adding new details about deaths from specific cardiovascular and respiratory causes (Pope et al., 2004b).
Significant associations were reported with deaths from specific cardiovascular diseases, particularly
ischemic heart disease, and a group of cardiac conditions including dysrhythmia, heart failure and cardiac
arrest (RR for cardiovascular mortality = 1.12, 95% CI 1.08-1.15 per 10 (ig/m3 PM25), but no PM
associations were found with respiratory mortality.
California Cancer Prevention Study: In a cohort of elderly people in 11 California counties (mean
age 73 years in 1983), an association was reported for long-term PM2 5 exposure with all-cause deaths
from 1973-1982 (RR= 1.04 [95% CI: 1.01-1.07] per 10 (ig/m3 PM2 5) (Enstrom, 2005). However, no
significant associations were reported with deaths in later time periods when PM2 5 levels had decreased
in the most polluted counties (1983-2002) (RR= 1.00 [95% CI: 0.98-1.02] per 10 (ig/m3 PM2 5). The
PM2 5 data were obtained from the EPA's Inhalation Particle Network (collected circa 1980), and the
locations represented a subset of data used in the 50-city ACS study (Pope et al., 1995). However, the use
of average values for California counties as exposure surrogates likely leads to significant exposure error,
as many California counties are large and quite topographically variable.
AHSMOG: In this analysis for the Seventh-Day Adventist cohort in California, a positive,
statistically significant, association with coronary heart disease mortality was reported for 92 deaths
among females (RR = 1.42 [95% CI: 1.06-1.90] per 10 (ig/m3 PM2 5), but not for 53 deaths among males
(RR = 0.90 [95% CI: 0.76-1.05] per 10 (ig/m3 PM2 5) (Chen et al., 2005a). Associations were strongest in
the subset of postmenopausal women (80 deaths; RR = 1.49 [95% CI: 1.17-1.89] per 10 (ig/m3 PM2 5).
The authors speculated that females may be more sensitive to air pollution-related effects, based on
differences between males and females in dosimetry and exposure.
U.S. Cystic Fibrosis cohort: A positive, but not statistically significant, association was reported
in this cohort (RR = 1.32 [95% CI: 0.91-1.93] per 10 (ig/m3 PM2 5) in a study that primarily focused on
evidence of exacerbation of respiratory symptoms (Goss et al., 2004). However, only 200 deaths had
occurred in the cohort of over 11,000 people (average age in cohort was 18.4 years), so the power of this
study to detect associations was low.
December 2008
7-105
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Women's Health Initiative (WHI) Study: This nationwide cohort study considered 65,893
postmenopausal women with no history of cardiovascular disease who lived in 36 U.S. metropolitan areas
from 1994 to 1998 (Miller et al., 2007b). The study had a median subject follow-up time of six years.
Miller and colleagues assessed each woman's exposure to air pollutants using the monitor located nearest
to their residence. Hazard ratios were estimated for the first cardiovascular event, adjusting for age, race
or ethnic group, smoking status, educational level, household income, body-mass index, and presence or
absence of diabetes, hypertension, or hypercholesterolemia. Overall, this study concludes that "long-term
exposure to fine particulate air pollution is associated with the incidence of cardiovascular disease and
death among postmenopausal women." In terms of effect size, the study found that each increase of 10
|ig/m3 of PM2 5 was associated with a 24% increase in the risk of a cardiovascular event (hazard ratio,
1.24 [95% CI: 1.09-1.41]) and a 76% increase in the risk of death from cardiovascular disease (hazard
ratio, 1.76 [95% CI: 1.25-2.47]). While this study found results confirmatory to the ACS and Six Cities
Study, it derives much larger relative risk estimates per |_ig/m3 PM2 5. This may be due to the fact that,
since the study included only women without pre-existing cardiovascular disease, it could potentially be a
healthier cohort population than that considered by the ACS and Six Cities Study. Indeed, the WHI Study
reported only 216 cardiovascular deaths in 349,643 women-years of followup, or a rate of 0.075% deaths
per year (Miller et al., 2007b), while the ACS Study reported that 10% of subjects died of cardiovascular
disease over a 16 year followup period, yielding a rate of 0.625% per year, or approximately 8 times the
cardiovascular mortality rate of the WHI population (Pope et al., 2004b). Thus, PM2 5 impacts may yield
higher relative risk estimates in the WHI population because the PM2 5 risk is being compared to a much
lower prevailing risk of cardiovascular death in this select study population.
The WHI study not only confirms the Six City Study and ACS Study associations with mortality in
yet another well characterized cohort with detailed individual-level information, it also has been able to
consider the individual medical records of the thousands of WHI subjects over the period of the study.
This has allowed the researchers to examine not only mortality, but also related morbidity in the form of
heart problems (cardiovascular events) experienced by the subjects during the study. As reported in this
paper, this examination confirmed that there is an increased risk of cardiovascular morbidity, as well (see
section 7.2.1). These morbidity co-associations with PM2 5 in the same population lend even greater
support to the biological plausibility of the air pollution-mortality associations found in this study.
Medicare Cohort Study: Using Medicare data, Eftim and co-authors (2008) have assessed the
association of PM2 5 with mortality for the same locations included in the Six City Study and the ACS
studies. For these locations, they estimated the chronic effects of PM2 5 on mortality for the period
2000-2002 using mortality data for cohorts of Medicare participants and average PM2 5 levels from
monitors in the same counties included in the two studies. Using aggregate counts of mortality by county
December 2008
7-106
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
for three age groups, they estimated mortality risk associated with air pollution adjusting for age and sex
and area-level covariates (education, income level, poverty, and employment), and controlled for potential
confounding by cigarette smoking by including standardized mortality ratios for lung cancer and COPD.
This study is, therefore, an ecological analysis, similar to past published cross-sectional analyses, in that
area-level covariates (education, income level, poverty, and employment) are employed as controlling
variables, since individual level information is not available from the Medicare database (other than age
and sex), which includes virtually all Americans aged 65 or greater. Exposures are also ecological in
nature, as central site data are used as indices of exposure. These results indicated that a 10 (ig/m3
increase in the yearly average PM2 5 concentration is associated with 10.9% (95% CI: 9.0-12.8) and with
20.8% (95% CI: 14.8-27.1) increases in all-cause mortality for the American Cancer Society and Harvard
Six Cities study counties, respectively. The estimates are somewhat higher than those reported by the
original investigators, and several possible explanations for this apparent increase are posited by the
authors, especially that this is an older population than the ACS cohort. Perhaps the most likely is that the
lack of personal confounder information (e.g., past personal smoking information) led to an insufficient
control for the effects of these other variables' effects on mortality, inflating the pollution effect estimates
somewhat, similar to what has been found in the ACS analyses when only ecological-level control
variables were included. The ability of the Eftim et al. (2008) study results to qualitatively replicate the
original individual-level cohort study (e.g., ACS and Six Cities Study) results suggests that past
ecological cross-sectional mortality study results may also provide useful insights into the nature of the
association, especially when used for consideration of time trends, or for comparisons of the relative
(rather than absolute) sizes of risks between different pollutants or PM components in health effects
associations.
Janes et al. (2007) use the same nationwide Medicare mortality data to examine the association
between monthly averages of fine particles (PM2 5) over the preceding 12 months and monthly mortality
rates in 113 U.S. counties from 2000 to 2002. They decompose the association between PM25 and
mortality into 2 components: (1) the association between "national trends" in PM2 5 and mortality; and
(2) the association between "local trends," defined as county-specific deviations from national trends.
This second component is posited to provide evidence as to whether counties having steeper declines in
PM2 5 also have steeper declines in mortality relative to their national trends. They report that the
exposure effect estimates are different at these 2 spatiotemporal scales, raising concerns about
confounding bias in these analyses. The authors assert that the association between trends in PM2 5 and
mortality at the national scale is more likely to be confounded than is the association between trends in
PM2 5 and mortality at the local scale and, if the association at the national scale is set aside, that there is
little evidence of an association between 12-month exposure to PM2 5 and mortality in this analysis.
December 2008
7-107
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
However, in response, Pope and Burnett (2007) point out that such use of long-term time trends as the
primary source of exposure variability has been avoided in most other air pollution epidemiology studies
because of such concerns about potential confounding of such time-trend associations.
By linking monitoring data to the U.S. Medicare system by county of residence, Zeger et al. (2007)
analyzed Medicare mortality records, comprising over 20 million enrollees in the 250 largest counties
during 2000-2002. The authors estimated log-linear regression models having as outcome the age-specific
county level mortality rates and, as the main predictor, the average PM2 5 pollution level in each county
during 2000. Area-level covariates were used to adjust for socio-economic status and smoking. The
authors reported results under several degrees of adjustment for spatial confounding and with
stratification into eastern, central and western U.S. counties. A 10 (ig/m3 increase in PM25 was associated
with a 7.6% increase in mortality (95% CI: 4.4-10.8). When adjusted for spatial confounding, the
estimated log-relative risks dropped by 50%. Zeger et al. (2007) found a stronger association in the
eastern counties than nationally, with no evidence of an association in western counties.
7.6.3. PM10-2.5
In the original analyses of the Six Cities and ACS cohort studies, no associations were found
between long-term exposure to PM10_2.5 and mortality, while the extended and follow-up analyses that are
discussed above did not evaluate potential associations with PM10_2 5. Two recent reports from the
AHSMOG and Veterans study cohorts have, however, provided some limited evidence for associations
between long-term exposure to PMi0.2.5 and mortality, as summarized below.
AHSMOG: As was found with fine particles, a positive association with coronary heart disease
mortality was reported for females (RR = 1.38 [95% CI: 0.97-1.95] per 10 (ig/m3 PMi0_2.5), but not for
males (RR = 0.92 [95% CI: 0.66-1.29] per 10 (ig/m3 PMio_2 5); associations were strongest in the subset of
postmenopausal women (80 deaths) (Chen et al., 2005a).
Veterans cohort: In this study (Lipfert et al., 2006a), a significant association was reported
between long-term exposure to PMi0.2.5 and total mortality in a single-pollutant model (RR = 1.07, 95%
CI 1.01-1.12 per 10 (ig/m3 PMi0.2.5). However, the association became negative and not statistically
significant in a model that included traffic density. As it would be expected that traffic would contribute to
the thoracic coarse particle concentrations, it is difficult to interpret the results of these multipollutant
analyses.
December 2008
7-108
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
7.6.4. PMio
The original analyses of the AHSMOG cohort study found positive associations between long-term
concentrations of PMi0 and 15-year mortality due to natural causes and lung cancer (Abbey et al., 1999).
McDonnell et al. (2000) reanalyzed these data and concluded that previously observed association of
long-term ambient PMi0 concentrations with mortality for males were best explained by a relationship of
mortality with the fine fraction of PMi0 rather than the coarse fraction of PMi0. Recent reports from the
AHSMOG study cohort, as well as the Nurses' Health Study, the U.S. Cystic Fibrosis cohort and a cohort
of women in Germany have, however, provided some evidence for associations between long-term
exposure to PMi0 and mortality among women, as summarized below.
AHSMOG: As was found with fine particles, a positive association with coronary heart disease
mortality was reported for females (RR = 1.22 [95% CI: 1.01-1.47] per 10 |_ig/nr' PMi0), but not for males
(RR = 0.94 [95% CI: 0.82-1.08] per 10 (ig/m3 PMi0); associations were strongest in the subset of
postmenopausal women (80 deaths) (Chen et al., 2005a).
Nurses' Health Study Cohort: The Nurses' Health Study (Puett et al., 2008) is an ongoing
prospective cohort study examining the relation of chronic particulate exposures with all-cause mortality
and incident and fatal coronary heart disease consisting of 66,250 female nurses in MSAs in the
northeastern region of the U.S. All cause mortality was statistically significantly associated with average
PMio exposures in the time period 3-48 months. The association was strongest with average PMi0
exposure in the 24 months prior to death (hazard ratio 1.16 [95% CI: 1.05-1.28]) and weakest with
exposure in the month prior to death (hazard ratio 1.04 [95% CI: 0.98-1.11]). The association with fatal
CHD occurred with the greatest magnitude with mean exposure in the 24 months prior to death (hazard
ratio 1.42 [95% CI: 1.11-1.81]).
U.S. Cystic Fibrosis Cohort: No clear significant association was reported in this cohort that
primarily focused on evidence of exacerbation of respiratory symptoms (Goss et al., 2004). However,
only 200 deaths had occurred in the cohort of over 11,000 people (average age in cohort was 18.4 years),
so the power of this study to detect associations was low.
German Cohort: The North Rhine-Westphalia State Environment Agency (LUANRW) initiated a
cohort of approximately 4800 women, and assessed whether long-term exposure to air pollution
originating from motorized traffic and industrial sources was associated with total and cause-specific
mortality (Gehring et al., 2006). They found that cardiopulmonary mortality was associated with PMi0
(RR= 1.52 [95% CI: 1.09-2.15] per 10 (ig/m3 PM10).
December 2008
7-109
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
7.6.5. Composition and Source-Oriented Analyses of PM
As discussed in the 2004 PM AQCD, only a very limited number of the chronic exposure cohort
studies have included direct measurements of chemical-specific constituents, or assessments of source-
oriented effects, of PM in their analyses. One exception is the Veterans Study, which looked at
associations with some constituents, and traffic.
Veterans Cohort: A recent reanalysis of the Veterans cohort data focused on exposure to traffic-
related air pollution (traffic density based on traffic flow rate data and road segment length) reported a
stronger relationship between mortality with long-term exposure to traffic than with PM2 5 mass (Lipfert
et al., 2000). A significant association was reported between total mortality and PM2 5 in single-pollutant
models (RR= 1.12 [95% CI: 1.04-1.20] per 10 (ig/m3 PM25). The authors observe that this risk estimate
is larger than results reported in a previous study of this cohort. In multipollutant models including traffic
density, the association with PM2 5 was reduced and lost statistical significance. Traffic emissions
contribute to PM2 5 so it would be expected that the two would be highly correlated, and, thus, these
multipollutant model results should be interpretted with caution. In a companion study, Lipfert et al.
(2006b) used data from EPA's fine particle speciation network, and reported findings for PM2 5 were
similar to those reported by Lipfert et al. (2006a). A positive association also was reported for mortality
with sulfates using the more recent data, but was not statistically significant. Using 2002 data from the
fine particle speciation network, significant associations were found between mortality and long-term
exposures to nitrates, EC, Ni and V, as well as traffic density and peak ozone concentrations. In
multipollutant models, associations with traffic density remained significant, as did nitrates, Ni and V in
some models.
Netherlands Study: Beelen et al. (2008a) studied the association between long-term exposure to
traffic-related air pollution and mortality in a Dutch cohort. They used data from an ongoing cohort study
on diet and cancer with 120,852 subjects who were followed from 1987 to 1996. Exposure to BS, N02,
S02, and PM2 5, as well as various exposure variables related to traffic, were estimated at the home
address. Cox analyses were conducted in the full cohort, adjusting for age, sex, smoking, and area-level
socioeconomic status. Traffic intensity on the nearest road was independently associated with mortality.
Relative risks (95% confidence intervals) for a 10 (ig/m3 increase in BS concentrations (difference
between 5th and 95th percentile) were 1.05 (95% CI: 1.00-1.11) for natural cause, 1.04 (95% CI:
0.95-1.13) for cardiovascular, 1.22 (95%CI: 0.99-1.50) for respiratory, 1.03 (95%CI: 0.88-1.20) for lung
cancer, and 1.04 (95% CI: 0.97-1.12) for mortality other than cardiovascular, respiratory, or lung cancer.
Results were similar for N02 and PM2 5, but no associations were found for S02. The authors concluded
that traffic-related air pollution and several traffic exposure variables were associated with mortality in
December 2008
7-110
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
the full cohort, although the relative risks were generally small. Associations between natural-cause and
respiratory mortality were statistically significant for N02 and BS. These results add to the evidence that
long-term exposure to traffic-related particulate ambient air pollution is associated with increased
mortality.
Given the general dearth of published source-oriented studies of the mortality impacts of long-term
PM exposure components, and given that the recent Medicare Cohort study now indicates that such
ecological cross-sectional studies can be useful for evaluating time trends and/or comparisons across
pollution components, it may well be that examining past cross-sectional studies comparing source-
oriented components of PM may be informative. In particular, Ozkaynak and Thurston (1987), utilized
the chemical speciation conducted in the Inhalable Particle (IP) Network to conduct a chemical
constituent and source-oriented evaluation on long-term PM exposure and mortality in the U.S. They
analyzed the 1980 U.S. vital statistics and available ambient air pollution data bases for sulfates and fine,
inhalable, and TSP mass. Using multiple regression analyses, they conducted a cross-sectional analysis of
the association between various particle measures and total mortality. Results from the various analyses
indicated the importance of considering particle size, composition, and source information in modeling of
particle pollution health effects. Of the independent mortality predictors considered, particle exposure
measures most related to the respirable fraction of the aerosols, such as fine particles and sulfates, were
most consistently and significantly associated with the reported SMSA-specific total annual mortality
rates. On the other hand, particle mass measures that included thoracic coarse particles (e.g., total
suspended particles and inhalable particles) were often found to be non-significant predictors of total
mortality. Furthermore, based on the application of fine particle source apportionment, particles from
industrial sources and from coal combustion were indicated to be more significant contributors to human
mortality than fine soil-derived particles.
7.6.6. Within-City Effects of PM Exposure
Much of the exposure gradient in the national-scale cohort studies was due to city-to-city
differences in regional air pollution, raising the possibility that some or all of the original PM-survival
associations may have been driven instead by city-to-city differences in some unknown (non-pollution)
confounder variable. This has been evaluated by two recent studies.
ACS, Los Angeles: To investigate this issue, a new analysis using ACS data focused on
neighborhood-to-neighborhood differences in urban air pollutants, using data from 23 PM2 5 monitoring
stations in the Los Angeles area, and applying interpolation methods to assign exposure levels to study
individuals (Jerrett et al., 2005b). This resulted in both improved exposure assessment and an increased
December 2008
7-111
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
focus on local sources of fine particle pollution. Significant associations between PM2 5 and mortality
from all causes and cardiopulmonary diseases were reported with the magnitude of the relative risks being
greater than those reported in previous assessments (after adjustment for potential confounders including
traffic, RR for cardiovascular diseases =1.17 [95% CI: 1.05-1.31], per 10 ug/m PM2,<; RR for ischemic
heart disease = 1.38 [95% CI: 1.11-1.72] per 10 pg/m3 PM2s). This indicates that city-to-city confounding
was not the cause of the associations found in the earlier ACS Cohort studies. The authors also suggest
that reducing exposure error can result in even stronger associations between PM2 5 and mortality than
generally observed in broader studies having less exposure detail.
WHI Study: This study also investigated the within- vs. between-city effects in its cities. As shown
in Figure 7-9, similar effects for both the within and between-city analyses demonstrate that this
association is not due to some other (non-pollution) confounder differing between the various cities,
strengthening confidence in the pollution-effect estimates.
A Overall Effect
12-
11-
10-
E
* B '
X v 9,
¦m Vi G
« fi
01 Q 7
O w 7-
w J2
o 3 6-
JX u
w £ C
s 5 5-
TJ 4-
lu 3-
2-
Q~| 1 nn 1111111»u 1111 n 11 |i n 11111111| 111 u| 1111 i;ii in |m 1 ii|
0 3 6 9 12 15 18 21 24 27 30
||llll|l|ll||M|ll|ll|||||ll|||l|lll|ll|ll||ll|l|T1lll|IHll|
0 3 6 9 12 15 18 21 24 27 30
PM2.5 HI™1)
B Between-City Effect
12-i
11-
10-
9-
E
o 1)
* s
X
,E
I
0"pill I |ll III | III ll| I HI l|l 1111)111 ll| III I l[llll I |ll III | HI H|
0 3 6 9 12 15 18 21 24 27 30
ITiiriiTnnii'iiiirn'inilTiiri'rmTiniri'iTtiHiniiim'iri
0 3 6 9 12 15 IS 21 24 27 30
7-
6-
5-
4-
3-
2-
1-
0 |i i in |m ii| in i hi in i |ii in |in n| im i|i im |iu min i n
0 3 6 9 12 15 18 21 24 27 30
li mii |iiiii|iiii i|niii |iiiii|iimi| inii|i mi |iiiii|ini ii
0 3 6 9 12 15 18 21 24 27 30
PM2.5 tng/m1)
Source: Miller et al. (2007b)
Figure 7-9. Plots of the relative risk of death from cardiovascular disease from the Women's
Health Initiative study displaying the between-city and within-city contributions to the
overall association between PM2.5 and cardiovascular mortality windows of exposure-
effects.
December 2008
7-112
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
7.6.7. Effects of Different Long-term Exposure Windows
The delay between changes in exposure and changes in health has important policy implications.
Schwartz et al. (2008) investigated this issue using an extended follow-up of the Harvard Six Cities Study.
Cox proportional hazards models were fit controlling for smoking, body mass index, and other covariates.
Penalized splines were fit in a flexible functional form to the concentration response to examine its shape,
and the degrees of freedom for the curve were selected based on Akaike's information criterion. They also
used model averaging as an alternative approach, where multiple models are fit explicitly and averaged,
weighted by their probability of being correct given the data. The lag relationship by model was averaged
across a range of unconstrained distributed lag models (i.e., same year, year prior, two years prior, etc.).
Results of the lag comparison are shown in Figure 7-10 indicating that the effects of changes in exposure
on mortality are seen within two years. The authors also noted that the concentration-response curve was
linear, clearly continuing below the level of the current U.S. air quality standard of 15 |ag/nr\
1.20
2 1.15
"O
Z 1.10
w
> 1.05
tr i.oo
0.95
Source: Schwartz et al. (2008)
Figure 7-10. The model-averaged estimated effect of a 10- pg/m3 increase in PM2.5 on all-cause
mortality at different lags (in years) between exposure and death. Each lag is
estimated independently of the others. Also shown are the pointwise 95% CIs for each
lag, based on jacknife estimates.
Roosli et al. (2005) took an alternative approach to determining the window over which the
mortality effects of long-term pollution exposures occurred. They fit the model shown in Figure 7-11
using k = 0.5 based on the Utah Steel Strike (Pope, 1989) and the Ireland coal ban study (Clancy et al.,
2002). They found that roughly 75% of health benefits are observed in the first 5 years, as shown in
12 3 4
Year before death
December 2008
7-113
DRAFT-DO NOT QUOTE OR CITE
-------
1 Table 7-9. This suggests that the most recent years of exposure are most important to mortality, consistent
2 with the findings of Schwartz et al. (2008).
io
o -
o
o ---
cr> -
ci
o
<33 _
° 1 1 1 1 1
2000 2002 2004 2006 2008
Time (years)
Source: Roosli et al. (2005)
Figure 7-11. Time course of relative risk of death after a sudden decrease in air pollution exposure
during the year 2000, assuming a steady state model (solid line) and a dynamic model
(bold dashed line). The thin dashed line refers to the reference scenario.
3 Puett et al. (2008) also compared different long-term lags, with exposure periods ranging from 1
4 month to 48 months prior to death. They found statistically significant associations with average PMi0
5 exposures in the time period 3-48 months prior to death, with the strongest associations in the 24 months
6 prior to death and the weakest with exposure in the 1 month prior to death. These results indicate a
7 developing coherence of the air pollution mortality literature, and the mortality risk benefits from
8 reducing air pollution would be expected within a few years of intervention.
December 2008
7-114
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Table 7-9. Distribution of the effect of a hypothetical reduction of 10 pg/m3 PM10 in 2000 on all-
cause mortality 2000-2009 in Switzerland.
Year 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Proportion of total effect (%) - 39.3 23.9 14.5 8.8 5.3 3.2 2.0 1.2 0.7 0.4
Relative risk (per 10 (jg/rn3 reduction in 1.0 0.9775 0.9863 0.9917 0.9950 0.9969 0.9981 0.9989 0.9993 0.9996 0.9997
PM10)
Relative risk and proportion of total effect in each year are shown, assuming a time constant k of 0.5
Source: Roosli et al. (2005)
7.6.8. Summary and Causal Determinations
The recent evidence is largely consistent with past studies, further supporting the evidence of
associations between long-term PM2 5 exposure and increased risk of human mortality in areas with mean
concentrations from 14 to 29 (ig/m3 (Figure 7-8). New evidence from the Six Cities cohort study shows a
relatively large risk estimate for reduced mortality risk with decreases in PM2 5 (Laden et al., 2006). The
results of new analyses from the Six Cities cohort and the ACS study in Los Angeles suggest that
previous and current studies may have underestimated the magnitude of the association (Jerrett et al.,
2005b). With regard to mortality by cause-of-death, the most recent ACS analysis (Pope et al., 2004b)
indicates that cardiac mortality primarily accounts for the total mortality association with PM2 5 among
adults, and not respiratory mortality. The recent WHI cohort study shows even higher cardiac risks
per |ig/m3 than found in the ACS study, but this is likely due to the fact that the study included only
women without pre-existing cardiovascular disease. Furthermore, the WHI study also considered within
vs. between city mortality, as well as morbidity co-associations with PM2 5 in the same population. The
first showed that the results are not due to between city confounding, and the morbidity analyses show the
coherence of the mortality association across health endpoints, supporting the biological plausibility of
the air pollution-mortality associations found in these studies.
The findings from a multiyear expert judgment study that comprehensively characterizes the size
and uncertainty in estimates of mortality reductions associated with decreases in PM2 5 in the U.S
provides additional support for an association between long-term exposure to PM2 5 and mortality (Roman
et al., 2008). This study applied state-of-the-art expert judgment elicitation techniques to develop
probabilistic uncertainty distributions that reflect the broader array of uncertainties in the concentration-
response relationship. This study followed best standard practices for expert elicitations based on the
body of literature accumulated over the past two decades, including: explicit criteria for expert selection,
a detailed interview protocol, briefing materials provided to experts in advance of the interview, and
workshops prior to and following the PM expert elicitation. The main goal of the protocol was to answer
December 2008
7-115
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
the following question: "What is your estimate of the true percent change in annual, all-cause mortality in
the adult U.S. population resulting from a permanent 1 (ig/m3 reduction in annual average ambient PM2 5
across the U.S.?" The resulting PM25 effect estimate distributions, elicited from 12 of the world's leading
experts on this issue, are shown in Figure 7-12. They indicate both larger central estimates of mortality
reductions for decreases in long-term PM2 5 exposure in the U.S. (averaging almost 1% per |ig/m3 PM2 5)
than reported (for example) by the ACS Study (i.e., 0.6% per (ig/m3 PM2 5 in (2002), and a wider
distribution of uncertainty by each expert than provided by any one of the PM2 5 epidemiologic studies.
However, a composite uncertainty range of the overall mean effect estimate (i.e., based upon all 12
experts' estimates, but not provided in Figure 7-12) would be much narrower, and closer to that derived
from the ACS study than indicated for any one expert shown in Figure 7-12.
t
o
a
s!
2.5
1.5
0.5
1
I
Group 1
Range of PMj 5 0
Concentration
Causality Likelihood
Expert
"JL.
I
Group 2
®i
44-
4-30 4-10 >10-30
99% 75% 99%
E L
4-10 >10-30 4-30 4-30 4-30 4-16 >16-30 4-7 >7-30 4-30 4-30 4-30 4-30 Pope Dockery
98% 98% 95% 95% 70% 35% 35% 100% 100% 99% 99% 95% 90% et al et aL
B DIG K F C J A H 2002 1993
Key: Closed circle = median; Open circle = mean; Box =• interquartite range; Solid line = 90% credible interval
Source: Roman et al. (2008)
Figure 7-12. Experts' mean effect estimates and uncertainty distributions for the PM2.5 mortality
concentration-response coefficient for a 1 |jg/m3 change in annual average PM2.5
In the 2004 PM AQCD, results from the ACS and Six Cities study analyses indicated that thoracic
coarse particles were not associated with mortality. The new findings from AHSMOG and Veterans cohort
studies provide limited evidence of associations between long-term exposure to PMi0.2.5 and mortality in
areas with mean concentrations from 16 to 25 |_ig/m3.
December 2008
7-116
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Overall, recent evidence supports the strong evidence reported in the 2004 PM AQCD that long-
term exposure to PM2 5 is associated with an increased risk of human mortality, though evidence is still
lacking to adequately characterize the association between PM10.2.5 and PM sources and/or components.
Collectively, the evidence is sufficient to conclude that the relationship between long-term PM2.5
exposures and mortality is likely to be causal. When looking at the cause of death, the strongest
evidence comes from mortality due to cardiovascular disease, with additional evidence supporting an
association between PM2 5 and lung cancer mortality (Figure 7-8). There is little new evidence that
supports an association between PM2 5 exposure and respiratory mortality (Figure 7-8), though the
existing evidence from the Harvard Six Cities and ACS studies show a strong relationship with
cardiopulmonary mortality (Figure 7-7). These findings are consistent and coherent with the evidence
from epidemiologic, human clinical and animal toxicological evidence for the effects of short-term
exposure on cardiovascular morbidity presented in Section 6.2., short- and long-term exposures on
respiratory morbidity (Sections 6.3. and 7.3, respectively), and infant mortality presented in Section 7.5.
Additionally, the evidence for short-term cardiovascular morbidity and short- and long-term respiratory
morbidity provide biological plausibility for mortality due to cardiovascular or respiratory disease. The
evidence for PM10 is suggestive of a causal relationship between long-term exposures and
mortality. The most recent evidence for this association is particularly strong for women. The evidence
for PM10-2.5 is inadequate to determine if a causal relationship exists between long-term exposures
and mortality.
December 2008
7-117 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Chapter 8. Public Health Impacts
This section addresses several issues relating to the broader public health impact from exposure to
ambient PM through a discussion on: (1) the shape of the concentration-response (C-R) relationship for
PM, with an evaluation of the new evidence available to assess a population threshold value for health
effects; and (2) the identification of subpopulations which may experience increased risks from PM
exposures, through either enhanced susceptibility (e.g., as a result of pre-existing disease, genetic factors,
age) and/or vulnerability associated with differential exposure (e.g., close proximity to sources,
activities).
8.1. Concentration-Response Relationship
An important consideration in characterizing the overall public health impacts associated with PM
exposure is whether the C-R relationship is linear across the full concentration range that is encountered
or if there are concentration ranges where there are departures from linearity (i.e., nonlinearity). The 2004
PM AQCD found, through the examination of multi-city studies (primarily those conducted using
NMMAPS data) that the linear model adequately represented the PM C-R relationship. In this ISA
additional studies have been identified that attempt to characterize the shape of the PM C-R curve. During
the evaluation of these studies particular interest is given to the shape of the C-R curve at and below the
current PMi0 and PM2 5 NAAQS daily level (PMi0: 150 (ig/m3 and PM2 5: 35 |_ig/m3) and annual level
(PM25: 15 (.ig/nr1).
In addition to examining the shape of the C-R relationship studies have also attempted to
identifying possible PM "thresholds" (i.e., levels which PM concentrations must exceed in order to elicit
a health response). An evaluation of multi-city studies in the 2004 PM AQCD found no evidence for the
presence of a threshold whereas, single-city studies did provide some suggestive evidence, but not in a
statistically clear manner. Overall, a multitude of factors have been identified that complicate the ability
to determine the shape of the PM C-R curve and the potential presence of a threshold including:
interindividual variability; additivity of pollutant-induced effects to naturally occurring background
disease processes; exposure error; response error; and low data density in the lower concentration range.
With consideration of these limitations, epidemiologic studies that examined the shape of the C-R curve
and the potential presence of a threshold for different exposure durations are presented below. The
discussion focuses on mortality effects associated with short- and long-term exposure to PM for which the
most in-depth analysis of the C-R relationship has occurred.
December 2008
8-1
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
8.1.1. Mortality Associated with Short-Term Exposure to PM
Although studies have consistently found that the C-R relationship between PM exposure and
mortality is linear and does not suggest the presence of a threshold, novel and more complex statistical
analyses continue to be developed to further analyze both the PM C-R curve and whether a threshold
exists. In an analysis of the PM-mortality C-R relationship for short-term exposure to PMi0, Daniels et al.
(2004) constructed three different models (i.e., (1) log-linear model, (2) spline model, and (3) threshold
model) to investigate the shape of the curve and whether a threshold exists. In their analysis the spline
model, which would allow for departures from linearity, showed a linear relationship without indicating a
threshold for both total (non-accidental) and cardiorespiratory mortality for each 10 (ig/m3 increase in
PMio, but the authors did find evidence for a threshold at 50 (ig/m3 for other-cause mortality (i.e., total
minus cardiorespiratory mortality). Further analysis of these models using AIC suggest that the log-linear
model is appropriate when describing the PMio-mortality relationship. However, it must be noted, as
stated by the HEI review committee, that AIC was not developed to assess scientific theories of etiology
and, therefore, the results obtained from this analysis, although consistent with the findings from previous
C-R relationship analyses, must be viewed with caution.
Schwartz (2004b) used a different technique to examine the C-R relationship, the inclusion of
indicator variables for days in which the PM10 concentration was between 15 and 25 (ig/m3, 25 and
34 (ig/m3, 35 and 44 (ig/m3, and above 45 (ig/m3. In this analysis the authors did not find any evidence for
deviations from linearity when combining estimates across 14 cities. Schwartz (2004b) did not analyze
city-specific thresholds, but Samoli et al. (2005), in the analysis of the C-R relationship in 22 European
cities observed heterogeneity in the shape of the C-R curve across cities. Therefore, although the
combined analysis conducted by Samoli et al. (2005) supports a log-linear association between PMi0 and
mortality, the heterogeneity observed between cities complicates the biological explanation for the
combined and city-specific results. Overall, the aforementioned studies all support the use of a
no-threshold log-linear model, but additional issues such as the influence of heterogeneity in estimates
between cities, and the effect of seasonal and regional differences in PM on the C-R relationship still
require further investigation.
8.1.2. Mortality Associated with Long-Term Exposure to PM
In addition to examining the C-R relationship between short-term exposure to PM and mortality,
one study conducted an analysis of the shape of the C-R relationship associated with long-term exposure
to PM. Schwartz et al. (2008) examined the C-R relationship between long-term exposure to PM2 5 and
December 2008
8-2
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
mortality using data from the Harvard Six Cities Study. The authors used two approaches both of which
involved Cox proportional hazards models. The first method used penalized splines, which fit a flexible
functional form that allowed for analysis of the C-R, while the second approach used Bayesian model
averaging (BMA) to combine the results of multiple models. In addition, the BMA approach used
piecewise linear functions to examine slope changes at various PM2.5 concentrations (i.e., 10, 15, 20, 25,
and 30 |_ig/m3). Using both approaches, the C-R curve was found to be indistinguishable from linear, and,
therefore, little evidence was observed to suggest that a threshold exists in the association between
long-term exposure to PM2.5 and the risk of death (Schwartz et al., 2008).
To date, the majority of studies that examined the C-R relationship between long-term exposure to
PM and mortality have assumed a no-threshold log-linear model; however, uncertainty still exists
surrounding the shape of the C-R curve. To further evaluate this uncertainty, EPA conducted an expert
elicitation (Roman et al., 2008) to develop probabilistic uncertainty distributions in an attempt to
characterize the array of uncertainties in the C-R relationship. During this process 12 experts were asked
to provide their judgment on the true shape of the C-R curve for annual average PM2.5 concentrations
ranging from 4 to 30 |ig/nr\ Eight of the 12 experts specified that the existing data is consistent with a
log-linear C-R curve. The remaining four experts all proposed non-linear functions that used two log-
linear segments. These experts suggested the use of a piece-wise function in order to account for the
uncertainty in mortality effects at low PM2 5 concentrations. Of the four experts that selected a non-linear
curve only one expert believed a threshold existed, even though the remaining experts agreed that,
collectively, the epidemiologic data did not provide evidence of a population threshold.
8.1.3. Summary of Concentration-Response Relationship
The examination of the PM C-R curve has primarily occurred in studies that have analyzed the
association between short- and long-term exposure to PM and mortality. These studies have used various
statistical methods, but overall have consistently found that a no-threshold log-linear model adequately
portrays the PM-mortality C-R relationship in multi-city analyses. At this time, uncertainty still exists
surrounding the PM-mortality C-R relationship on a city-to-city basis due to heterogeneity in the shape of
the C-R curve across cities.
8.2. Potentially Susceptible and Vulnerable Subpopulations
Interindividual variation in human responses to air pollutants indicates that not all individuals
exposed to pollutants respond similarly. That is, some subpopulations are at increased risk to the
December 2008
8-3
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
detrimental effects of pollutant exposure. The NAAQS are intended to provide an adequate margin of
safety for both general populations and sensitive subpopulations, or those subgroups potentially at
increased risk for ambient air pollution health effects. For the purposes of this document, a susceptible
subpopulation is defined as one that might exhibit an adverse health effect to a pollutant at concentrations
lower than those needed to elicit the same response in the general population or elicit a more adverse
effect to the same concentration. A vulnerable subpopulation is one that might be differentially exposed to
higher concentrations of a pollutant than the general population, regardless of the health outcome. The
previous review of the PM NAAQS identified certain groups within the population that may be more
susceptible to the effects of PM exposure, including infants, older adults, asthmatics, individuals with
COPD or cardiovascular disease, diabetics, and individuals with certain genetic polymorphisms. Other
subgroups considered to be somewhat vulnerable in the previous review include individuals that
encompass particular socioeconomic status (SES) groups and education levels; exercising individuals;
and those living near roadways. Table 8-1 provides an overview of the characteristics that contribute to
susceptible/vulnerable subpopulations, which have been observed in the examination of the NAAQS for
all criteria pollutants. Those characteristics of susceptible/vulnerable subpopulations exposed specifically
to PM, as mentioned in the literature that encompasses this ISA, are discussed below.
Table 8-1. Characteristics of susceptible/vulnerable subpopulations.
Susceptibility Characteristics1
Vulnerability Characteristics2
Age: Children, Older Adults (65+
Education Level
Infants: Premature, Low Birth Weight
Air Conditioning Use
Pregnancy
Proximity to Roadways
Birth Defects
Geographic Location (West vs. East)
Gender
Level of Exercise
Race/Ethnicity
Work Environment (e.g., outdoor workers)
Genetic Factors
Socioeconomic Status
Pre-existing disease: Obesity, Diabetes, Respiratory diseases (e.g., asthma),
Cardiovascular diseases
Nutritional status
1 Susceptible (i.e., intrinsic) refers to biological characteristics of an individual, which can include life stage, genetics, and pre-existing disease.
2 Vulnerable (i.e., extrinsic) refers to non-biological variables associated with an individual that can result in a health effect.
December 2008
8-4
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
8.2.1. Susceptibility Characteristics
8.2.1.1. Age
The 2004 PM AQCD found evidence that age modifies the health effects associated with exposure
to PM. Depending on the health effect under investigation, children and older adults (i.e., > 65 years)
have been identified as the two most susceptible subpopulations. Most studies that age-stratify results
have typically reported associations between PM and respiratory-related health effects for children,
specifically asthma, and associations for cardiovascular-related disease in older adults (U.S. EPA, 2004).
As the demographics of the U.S. population shift over the next 20 years with a larger percentage of the
population (i.e., 13% of the population in 2011 and a projected 20% in 2030) encompassing individuals
over the age of 65, an increase in the number of PM-related health effects could occur
(U.S. Census Bureau, 2000).
The recent epidemiologic, human clinical, and toxicological literature has examined the role of age
on the health effects observed upon exposure to PM. Overall, the results from epidemiologic studies vary,
depending on the health outcome of interest, in regards to which ages are most susceptible to PM
exposure. Studies that examined the effect of short-term exposure to PM on cardiovascular morbidity,
specifically, hospital admissions were found to exhibit a greater degree of variability in the estimates
reported between studies depending on the study location (i.e., U.S. or Europe). The majority of studies
conducted in Europe that present age-stratified results found that age (i.e., > 65) modifies the PM risk
associated with cardiovascular hospital admissions. Le Terte et al. (2002b) found in the European
APHEA2 study that the excess risk of hospitalization for IHD attributable to PMi0 was approximately
twice as large in patients aged > 65 years as compared to those aged < 65 years. Barnett et al. (2006)
analyzed data from several cities across Australia and New Zealand and found that the excess risk of
hospitalizations for cardiac diseases, congestive heart failure, IHD, MI, and all CVD was greater among
patients aged > 65 as compared to those individuals <65 years upon exposure to PM2 5. The French PSAS
program also found that the excess risk of hospitalization for all CVD, cardiac diseases, and IHD
attributable to PMi0, PM2 5, andPMio-2 5 was consistently greater among patients aged > 65 years than in
all ages combined (Host et al., 2007; Larrieu et al., 2007) In many cases, the PMi0 effects were only
significant among older adults; however, formal tests of heterogeneity were commonly not reported in
these studies. A detailed evaluation of the U.S.-based studies that examined the association between
short-term exposure to PM and cardiovascular morbidity did not find consistent results across studies.
Only Zanobetti and Schwartz (2002) in their analysis of the effect of PMi0 on cardiac hospital admissions
among Medicare beneficiaries in 4 cities found that age (i.e., >75 compared to individuals 65-75)
modified the PM risk. In addition, two studies that examined the association between short-term exposure
December 2008
8-5
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
to PM10 and hospitalizations for MI found no evidence of effect modification by age (Zanobetti and
Schwartz, 2005) among Medixare beneficiaries in 21 cities and Metzger et al. 2004 in the Atlanta-based
SOPHIA study).
Although epidemiologic studies did not consistently find that age modified PM risk estimates for
cardiovascular morbidity, data from human clinical studies provide support for an increase in
cardiovascular effect in older adults. Human clinical studies that exposed individuals to CAPs have found
evidence of increased cardiovascular responses in older subjects. Devlin et al. (2003) found that older
subjects exposed to fine particulate CAPs experienced significant decreases in HRV (both in time and
frequency) immediately following exposure, when compared to data in healthy young subjects. In
addition, Gong et al. (2004b) found that older subjects showed significant decreases in HRV when
exposed to CAPs, but this study did not compare the response in older subjects to those elicited by young,
healthy individuals. Although cardiovascular morbidity epidemiologic studies were unable to consistently
suggest that older adults are more susceptible to PM, decreased HRV in older individuals has been shown
to predict increased risk of cardiovascular morbidity. (Brook et al., 2004).
Animal models have also been developed that mimic the physiologic conditions associated with
older individuals in order to examine PM-related health effects. For example, Nadziejko et al. (Nadziejko
et al., 2004) observed arrhythmias in older, but not younger, rats exposed to fine CAPs. In addition,
another study (Tankersley et al., 2004) that used a mouse model of terminal senescence observed various
cardiovascular-related responses including: altered baseline autonomic tone in response to carbon black
exposure which may affect the quality and severity of cardiovascular responses (Tankersley et al., 2007).
Reductions in cardiac fractional shortening and significant pulmonary vascular congestion upon exposure
to carbon black were also reported in old mice (Tankersley et al., 2008). Overall, these studies do provide
some biological plausibility for the increase in cardiovascular effects in older adults observed in the
human clinical studies.
Epidemiologic studies that examined the association between exposure to PM and respiratory
morbidity, and whether age modifies the effects observed, found evidence that supports the findings of the
2004 PM AQCD, which suggested that children are more susceptible to respiratory-related health effects
(Barnett et al., 2005; Mar et al., 2004; Peel et al., 2005). Mar et al. (2004) found that children exposed to
various PM size fractions were at an increased risk of developing lower respiratory symptoms in Spokane,
Washington. In addition, Peel et al. (2005) and Barnett et al. (2005) both observed that children were
more susceptible to respiratory-related hospital admissions upon exposure to PMi0. The literature has not
consistently found an association between short-term exposure to PM and respiratory-related health
effects in older adults, but some studies have reported an increase in respiratory hospital admissions
(Andersen et al., 2007b; Fung et al., 2006). The results from dosimetry studies have shown a depression
December 2008
8-6
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
of PM clearance throughout the respiratory tract with increasing age from young to adulthood in humans
and laboratory animals, which could lend support to older adults also being susceptible to PM-related
respiratory health effects although this has not yet been supported by the human clinical or epidemiologic
literature (Section 4.3.4.1).
The new epidemiologic studies that examine the effect of short-term exposure to PM on mortality
have found that individuals over the age of 65 are more susceptible to all-cause (non-accidental) mortality
upon exposure to both PMi0 (Zeka et al., 2006a) and PM2 5 (Franklin et al., 2007; Ostro et al., 2006),
which is consistent with the findings of the 2004 PM AQCD. However, Ostro et al. (2006) only observed
a slight increase in mortality for older adults, but they did observe that the inclusion of gaseous
copollutants (i.e., CO and N02) in the older adults model did not affect the PM2 5 coefficient, unlike the
all-age model whose effect estimate was attenuated upon the inclusion of CO and N02 (Ostro et al.,
2006). Therefore, the results obtained by Ostro et al. (2006) further suggest that older adults are more
susceptible to PM exposures even though the overall effect estimate does not differ significantly from the
estimate obtained from the all-ages model. Studies that examined the effects of long-term exposure to PM
have found results contradictory to those obtained for mortality attributed to short-term exposures.
Villeneuve et al. (2002) found that individuals < 60 f had the greatest risk of PM-related mortality;
whereas, Lipfert et al. (2002) observed evidence that suggested that PM2 5 disproportionately affects
individuals < 65 years while individuals > 65 years are more susceptible to PM10.
8.2.1.2. Pregnancy
Pregnant women may be a susceptible subpopulation primarily due to the potential effect of
environmental contaminants on the developing fetus. In the case of exposure to PM, adverse health
effects in the offspring are mediated by a health response in the pregnant woman. Fedulov et al. (2008)
used an animal model to examine the effect of DEPs along with an immunologically "inert" particle
(titanium dioxide [Ti02]) on pregnant mice. The authors found that pregnant mice exhibited a local and
systemic inflammatory response when exposed to both DEP and Ti02, which was not observed in control,
non-pregnant mice. In addition, the offspring of exposed pregnant mice developed AHR and allergic
inflammation. The importance of this finding is that an inflammatory response leads to the differential
activation of multiple genes involved in immune response and regulation, cell metabolism, and
proliferation (2008).
In an additional study Hamada et al. (2007) also observed allergic responses in the offspring of
dams exposed to ROFA prior to delivery. The offspring responded to OVA immunization and aerosol
challenge with AHR and increased antigen-specific IgE and IgGl antibodies. Overall, these studies
December 2008
8-7
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
suggest that exposure to PM and even relatively inert particles during pregnancy can potentially lead to
increased allergic susceptibility in offspring, specifically, asthma.
8.2.1.3. Gender
Although the 2004 PM AQCD did not find consistent evidence to support a difference in health
effects by gender, there does appear to be gender differences in the localization of particles upon
deposition and the deposition rate (U.S. EPA, 2004). The recent epidemiologic studies that examine the
association between exposure to PM and mortality and morbidity provide inconclusive results as to
whether PM disproportionately affects males or females. Both Zanobetti and Schwartz (2005) and
Wellenius et al. (2006b) did not find gender to be a significant effect modifier of the risk estimates
associated with cardiovascular hospital admissions, but Pope et al. (2006) did observe a slightly larger,
non-significant, association between PM2 5 and hospitalization for acute IHD events in males. In the
examination of the association between short-term exposure to PMi0 and respiratory hospitalization,
Luginaah et al. used both a time-series and case-crossover design and found similar effects for both
males and females. Boezen et al. (2005) did observe differential effects for males and females depending
on the endpoint examined (i.e., slightly larger association for upper respiratory symptoms in males and
cough in females) during their analysis of the PMio-respiratory morbidity relationship. However, the
authors hypothesized that these differences could potentially be due to the differential daily exposure to
traffic exhaust experienced by men compared to women. Variable results were also observed in those
studies that examined the effect of short-term exposure to PM on mortality. Epidemiologic studies that
examined the relationship between short-term exposure to PM10 and mortality did not observe any effect
modification when stratifying by gender (Zeka et al., 2006a), while some PM2 5 studies (Franklin et al.,
2007; Ostro et al., 2006) did observe slightly larger, but non-significant, estimates in females compared to
males. Similar analyses were also conducted when examining the association between long-term exposure
to PM and mortality, but unlike the results presented above for short-term exposure studies, each study
consistently found that PM2 5 and/or PMi0 mortality risk estimates for females were slightly larger,
although not significantly so, than those for males (Chen et al., 2005a; Naess et al., 2007a; Zanobetti and
Schwartz, 2007).
8.2.1.4. Race/Ethnicity
Epidemiologic studies that examined the effect of race and ethnicity on morbidity and mortality
obtained largely inconsistent results. Wellenius et al. (2006b) and Zanobetti and Schwartz (2005) both
observed that race did not significantly modify the association between short-term exposure to PMi0 and
December 2008
8-8
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
CHF hospitalizations and MI hospitalizations, respectively. In the analysis of the PM10-mortality
relationship, Zeka et al. (2006a) did not observe any difference in mortality effect estimates when
stratifying by race (i.e., black and white) upon short-term exposure to PM10. However, Ostro et al. (2006)
when performing a similar analysis, but also including ethnicity, observed a positive and significant effect
for whites and Hispanics, but not for blacks in response to short-term exposure to PM25. An additional
analysis performed by Ostro et al. (2006) using PM2 5 and various PM2 5 species as the pollutants of
interest, also observed a significant association between mortality, specifically cardiovascular mortality,
and Hispanic ethnicity (Ostro et al., 2008). Overall, although significant associations have been observed
between various PM size fractions and race and/or ethnicity it remains unclear if either increases the
susceptibility of individuals to PM-related health effects.
8.2.1.5. Gene-Environment Interaction
A consensus now exists that gene-environment interactions merit serious consideration during the
examination of the relationship between ambient exposures to air pollutants and the development of
health effects (Gilliland et al., 1999; Kauffmann, 2004). Inter-individual variation in human responses to
air pollutants suggests that some subpopulations are at increased risk of detrimental effects due to
pollutant exposure, and it has become clear that the genetic makeup of an individual is an important
susceptibility factor (Kleeberger and Ohtsuka, 2005). Gene-environment interactions can result in health
effects due to either: (1) genetic polymorphisms, which result in the lack of a protein or a change that
makes a dysfunctional protein that is needed to maintain homeostasis in the body (e.g., scavenging of
ROS by glutathione-S-transferase [GST] genes), or (2) genetic damage in response to an exposure which
potentially leads to a health response (e.g., formation of benzo [a] pyrene DNA adducts in response to PM
exposure). To date, the majority of studies examine gene-environment interactions due to genetic
polymorphisms. In order to establish useful links between polymorphisms in candidate genes and adverse
health effects several criteria must be satisfied: (1) the product of the candidate gene must be significantly
involved in the pathogenesis of the adverse effect of interest; (2) polymorphisms in the gene must produce
a functional change in either the protein product or in the level of expression of the protein; and (3) the
issue of confounding by other environmental exposures must be carefully considered (U.S. EPA, 2008d).
It has been hypothesized that the cardiovascular and respiratory health effects that occur in
response to short-term exposure to PM are mediated by oxidative stress (see Section 4.3). Research has
examined this hypothesis by primarily focusing on the GST genes because they have common,
functionally important polymorphic alleles that significantly affect antioxidant defense function in the
lung (e.g., homozygosity for the null allele at the GSTM1 and GSTT1 loci, homozygosity for the A105G
allele at the GSTP1 locus). Exposure to radicals and oxidants in air pollution leads to a cascade of events,
December 2008
8-9
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
which can result in a reduction in glutathione (GSH), and an increase in the transcription of GSTs.
Individuals with genotypes that result in reduced or absent enzymatic activity are likely to have reduced
antioxidant defenses and potentially increased susceptibility to inhaled oxidants and radicals.
Numerous studies have examined the role of genetic polymorphisms on PM-related cardiovascular
health effects using the Normative Aging Study cohort. Schwartz et al. (2005b) and Chahine et al. (2007)
both found that individuals with null GSTM1 alleles had a larger decrease in HRV upon exposure to PM2.5
compared to individuals with at least one allele. Polymorphisms in the heme oxygenase-1 (HO-1)
promoter resulted in different responses in HRV upon exposure to PM2 5 depending on whether the
individual had the long or short repeat polymorphism (only those individuals with the long repeat
polymorphism had a decline in HRV) (Chahine et al., 2007). These results taken together suggest that
individuals with null alleles or specific polymorphisms in genes that mediate the antioxidant response to
oxidative stress are more susceptible to PM exposure. However, in some cases genetic polymorphisms
may actually reduce an individual's susceptibility to PM-related health effects. For example, Park et al.
(2006) found that individuals with 2 hemochromatosis (HFE) polymorphisms, which result in an increase
in iron uptake, had smaller reductions in HRV upon exposure to PM2 5.
Additional studies have also examined whether genetic polymorphisms increase the susceptibility
of individuals to respiratory morbidity in response to PM exposure. Gilliland et al. (2004) examined the
effect of allergens and DEPs on individuals with either null genotypes for GSTM1 and GSTT1 or GSTP1
codon 105 variants. They found that individuals with the GSTM1 null or the GSTP1 1105 wildtype
genotypes were more susceptible to allergic inflammation upon exposure to allergen and DEPs.
Although some of the aforementioned studies have observed associations between exposure to PM
and various genetic polymorphisms, multiple factors besides the specific polymorphism under
investigation may be causing the response observed. Overall, additional research is required to further
substantiate the influence of gene-environment interactions on susceptibility to health effects in response
to PM exposure.
8.2.1.6. Pre-Existing Disease
In 2004, the National Research Council (NRC) published a report that emphasized the need to
evaluate the effect of air pollution on susceptible groups, including those with respiratory illnesses and
cardiovascular diseases (NRC, 2004). The 2004 PM AQCD found epidemiologic evidence suggesting that
individuals with pre-existing heart and lung diseases, and diabetes may be more susceptible to PM
exposure. In addition, toxicological studies that used animal models of cardiopulmonary diseases and
heightened allergic sensitivity also found evidence of enhanced susceptibility. Since the previous PM
AQCD epidemiologic, toxicological, and human clinical studies have constructed models or directly
December 2008
8-10
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
examined the effect of PM on individuals with pre-existing diseases to identify whether exposure to PM
disproportionately effects certain subpopulations. A significant percentage of the U.S. population has
cardiovascular diseases and respiratory illnesses (see Table 8-2). In addition, diabetes and obesity, two
conditions linked to chronic inflammation, have recently been found to potentially facilitate PM-mediated
health effects. The large prevalence of diabetes in the U.S. population and the increasing percent of
individuals defined as overweight or obese (BMI > 25.0) (56%-65% between NHANES III and NHANES
[1999-2002)]), in combination with those individuals defined as having a pre-existing cardiovascular or
respiratory disease, constitute an extremely large percent of the U.S. population that may be susceptible to
PM-related health effects (NCHS, 2007).
8.2.1.7. Cardiovascular Diseases
The effect of underlying cardiovascular diseases on PM-related health effects was also examined
using epidemiologic and human clinical studies, along with toxicological studies that use models to
mimic the physiologic conditions associated with various cardiovascular diseases (e.g., MI, angina, and
atherosclerosis). Epidemiologic studies have observed cardiovascular health outcomes in individuals with
underlying cardiovascular diseases upon exposure to PM. An increase in risk estimates for associations
between PMi0 and mortality have been observed in individuals with underlying stroke (Zeka et al., 2006b)
and congestive heart failure (Bateson and Schwartz, 2004). In addition, studies that have focused on
morbidity outcomes have found an increase in emergency department (ED) visits for individuals with
various cardiovascular conditions. Metzger et al. (2007), in Atlanta, observed that exposure to PM10
resulted in an increase in ED visits for arrhythmias and CHF in individuals with underlying hypertension,
along with an increase in IHD ED visits for individuals with pre-existing arrhythmia; however, CHF did
not contribute to an increase in IHD ED visits. An additional study conducted in Atlanta, Peel et al. (Peel
et al., 2007), presented results consistent with those reported by Metzger et al. (2007) - secondary
hypertension increased the risk of CHF and arrhythmia ED visits upon exposure to PMi0. In contrast,
Pope et al. (2006) observed no association between secondary hypertension and IHD ED visits in Utah,
but they did find an increase in hospital admissions for acute IHD in individuals with underlying CHF
with exposure to PM2 5. Two additional studies also observed no effect modification during an analysis of
underlying cardiovascular diseases. Zanobetti and Schwartz (2005) did not find an increase in MI hospital
admissions for exposure to PMi0 in individuals with secondary atrial fibrillation or CHF in a cohort of
more than 300,000 hospital admissions. Wellenius et al. (2006b) also found no effect modification
between CHF hospital admissions and acute MI or arrhythmia in 7 U.S. cities in response to exposure to
PMio, such asHRV, . When looking at specific cardiac measurements, such as HRV upon exposure to
PM2 5, pre-existing cardiovascular conditions (i.e., hypertension and IHD) have been found to contribute
December 2008
8-11
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
to a reduction in the HF parameter (Park et al., 2005b). Overall, the epidemiologic evidence reports
inconsistent findings regarding the effect of pre-existing cardiovascular diseases on cardiovascular
hospital admissions and ED visits in response to exposure to PM.
Human clinical and toxicological studies have attempted to decipher the role of underlying
cardiovascular conditions on PM-related health effects by either designing studies that include subjects
with pre-existing cardiovascular diseases or employing animal models of a specific cardiovascular
disease. The effects of PM on rats that experienced a MI were analyzed by Anselme et al. (2007) and
Wellenius et al. (2006b) using two different models. Wellenius et al. (2006b) found, using the
post-myocardium sensitivity model (acute MI), that exposure to CAPs decreased spontaneous
supraventricular arrhythmias. In contrast, the MI model of chronic heart failure (rats that experienced an
MI 3 months prior to exposure), demonstrated a prominent increase in the incidence of premature
ventricular contraction when exposed to DE and this response was not observed in healthy rats (2007).
The discrepancies in effects could be due to differences in MI model or exposure atmosphere.
Wellenius et al. (2003) examined the effects of PM on angina using myocardial ischemic
preconditioning in dogs, which mimics the effects associated with IHD. In this study heart rate variability
was examined in response to exposure to fine particulate CAPs. The authors found that exposure to fine
particulate CAPs via tracheostomy prior to preconditioning increased integrated ST-segment elevation,
indicating ischemia.
The majority of the toxicology studies that examined the association between PM and pre-existing
cardiovascular diseases used a model of oxidative stress, the ApoE" "mouse model. Each of these studies
observed physiologic alterations in response to both short- and long-term exposure to various PM size
fractions. Campen et al. (2005; 2006) both examined the effects of short-term exposure to PM on ApoE""
mice. Campen et al. (2005) found that DE induced dramatic bradycardia and T-wave depression, while
Campen et al. (2006) found that whole gasoline emissions induced T-wave alterations.
In analyses that focused on the health effects associated with long-term exposure to PM on ApoE" "
mice, relatively consistent physiologic effects were observed across studies. Araujo et al. (2008) in a
exposed mice to ultrafine CAPs and found that long-term exposure to PM enhanced the size of aortic
lesions. Similarly, Chen and Nadziejko (2005) and Sun et al. (2005; 2008) exposed mice to PM2 5 CAPs
with the same results. Additional long-term exposure studies observed a decreasing trend in heart rate,
physical activity, and temperature along with biphasic responses in HRV (SDNN and rMSSD) (Hwang et
al., 2005) upon exposure to CAPs.
Human clinical studies have also examined the effect of pre-existing diseases on the health effects
associated with exposure to PM. Mills et al. (2007; 2008) investigated the effects of dilute diesel-exhaust,
and fine and ultrafine CAPs, respectively, on subjects with chronic heart disease (CHD). Exposure to
December 2008
8-12
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
dilute diesel-exhaust was found to promote exercise-induced myocardial ischemia and inhibit endogenous
fibrinolytic capacity (Mills et al., 2007), while fine and ultrafine CAPs, which were low in combustion
products, were not found to exhibit any significant effects on vascular function. An additional study
conducted by Carlsten et al. (2008), which also examined cardiovascular effects, found that DE did not
elicit any prothrombotic effects in subjects with metabolic syndrome.
8.2.1.8. Respiratory Illnesses
Investigators have examined the effect of pre-existing respiratory illnesses on multiple health
outcomes (e.g., mortality, asthma symptoms, congestive heart failure, etc.) in response to exposure to
ambient levels of PM. In some cases animal models have been developed and/or human clinical studies
conducted in order to substantiate the results obtained from epidemiologic studies. Unlike the analyses
conducted for pre-existing cardiovascular diseases, which examined the progression of only
cardiovascular conditions in response to PM exposure, the studies that examined the effect of pre-existing
respiratory diseases on PM-related health effects examined both respiratory and cardiovascular health
outcomes.
The majority of epidemiologic studies examined the effect of various pre-existing respiratory
illnesses on PM-related health effects, specifically in individuals with asthma or COPD. Although each
epidemiologic study was conducted in a different location, using (in some cases) different averaging times
and/or lags, an increase in respiratory effects in individuals with pre-existing asthma was consistently
observed in response to short-term exposure to size-fractionated PM. Individuals with pre-existing asthma
were found to have an increase in: medication use (Rabinovitch et al., 2006), respiratory symptoms
(i.e., asthma symptoms, cough, shortness of breath, and chest tightness (Gent et al., 2003), and asthma
symptoms (Delfino et al., 2002; Delfino et al., 2003a) when exposed to PM2 5; and morning symptoms
(Mortimer et al., 2002); and asthma attacks (Desqueyroux et al., 2002a) when exposed to PM10. The
results of epidemiologic studies that focused on individuals with pre-existing COPD are less consistent
than those reported for studies that examined the effect of short-term exposure to PM on individuals with
pre-existing asthma. Both Silkoff et al. (2005) and (Desqueyroux et al., 2002b) did not find an increase in
the exacerbation of COPD in response to short-term exposure to PM2 5. However, Trenga et al. (2006) and
Lagorio et al. (2006) did observe declines in lung function FEVi and FEVi and FVC, respectively in
response to PMi0 and/or PM2 5.
Collectively, human clinical studies that examined the health effects associated with pre-existing
respiratory diseases (i.e., asthma and COPD) did not find an increase in respiratory effects upon exposure
to PM. Gong et al. (2003a; Gong et al., 2008) found that healthy and asthmatic subjects exposed to fine
and ultrafine CAPs, exhibited similar respiratory responses. In addition, Gong et al. (2004a) found no
December 2008
8-13
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
significant difference in respiratory effects between healthy and COPD inflicted individuals upon
exposure to fine CAPs. However, the results from dosimetry studies, which have shown that mucociliary
clearance is impaired in patients with COPD relative to healthy controls, suggest that individuals with
preexisting COPD are potentially at a greater risk of COPD exacerbations (see Section 4.3.4.3), but this
has not yet been confirmed in human clinical or epidemiologic studies
Toxicological studies have examined the effect of pre-existing allergy on PM-related health effects
through the use of the ovalbumin-induced allergic airway disease model. Morishita et al. (2004) used this
model to assess the health effects of PM2 5 components. Using CAPs from Detroit, an area with pediatric
asthma rates three times the national average, rats with induced allergic airway disease were found to
preferentially retain PM derived from identified local combustion sources in association with eosinophil
influx and BALF protein content after a short-term exposure (Morishita et al., 2004). These findings
suggest that individuals with allergic airways conditions are more susceptible to allergic airways
responses upon exposure to PM2 5, which may be attributed to increased pulmonary deposition and
localization of particles in the respiratory tract (Morishita et al., 2004). This conclusion is supported by a
series of human clinical studies which have shown that exposure to DEPs increases the allergic
inflammatory response in atopic individuals (Bastain et al., 2003; Diaz-Sanchez et al., 1997; Nordenhall
et al., 2001).
Of the studies that focused on the association between short- and long-term exposure to PM and
mortality, only a few further refined their analysis to examine the effect of pre-existing respiratory
illnesses on the PM-mortality relationship. Using different pre-existing respiratory illnesses, Zeka et al.
(2006a) (examined pneumonia) and De Leon et al. (2003) (all respiratory illesses) found that short-term
exposure to PMi0 along with: (1) pneumonia increased the risk of non-accidental mortality, and (2)
respiratory illnesses increased the risk of circulatory mortality, respectively. Additionally, Zanobetti et al.
(2008) observed an association between long-term exposure to PMi0 and mortality in individuals that had
previously been hospitalized for COPD.
8.2.1.9. Respiratory Contributions to CV Effects
Although the majority of health effects observed in individuals with pre-existing respiratory
illnesses were associated with respiratory illness exacerbations, studies also examined whether underlying
respiratory illnesses can lead to cardiovascular effects upon PM exposure. The epidemiologic studies that
examined this association did not observe an increase in hospital admission or emergency department
visits for a variety of cardiovascular effects (e.g., ischemic heart disease, arrhythmias, congestive heart
failure, myocardial infarction) for individuals with underlying respiratory infection (Wellenius et al.,
2006b), pneumonia (Zanobetti and Schwartz, 2005), and COPD (Peel et al., 2007; Wellenius et al.,
December 2008
8-14
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
2005a). However, human clinical and toxicological studies did observe some cardiovascular effects in
individuals with pre-existing respiratory illnesses. Gong et al. (2003a) observed signs of acute responses
in the cardiovascular system, systemic circulation, and central-airway cell populations in asthmatic
individuals exposure to fine CAPs. Batalha et al. (2002), using a chronic bronchitis animal model, found
that the pulmonary artery lumen-to-wall ratio was decreased in chronic bronchitis rats exposed to both
filtered air and CAPs. Normal rats were also found to have a reduced pulmonary artery lumen-to-wall
ratio, but only when exposed to PM2 5 CAPs. Overall, it is unclear if underlying respiratory illnesses result
in increased susceptibility to cardiovascular health effects.
8.2.1.10. Inflammatory Conditions: Diabetes and Obesity
An increasing number of studies have been conducted since the 2004 PM AQCD to examine the
effects of short-term exposure to PM on individuals with chronic inflammatory conditions, such as
diabetes and obesity. As the percent of the population with each condition (i.e. diabetes and obesity)
continues to increase these subpopulations could represent a large number of individuals that are
susceptible to the health effects associated with exposure to PM. Although a few studies have found that
individuals with diabetes are at increased risk of mortality upon exposure to PMi0 (Goldberg et al., 2006;
Zeka et al., 2006a), the majority of the literature focuses on cardiovascular health effects in these
individuals.
It has been hypothesized that the systemic inflammatory cascade leads to an increase in
cardiovascular risk (Dubowsky et al., 2006). To examine the potential increase in risk associated with
short-term exposure to PM10 in diabetics, human clinical studies have been conducted to determine the
physiologic changes that occur in individuals with diabetes in response to PM exposure. These studies
examine both changes in inflammatory markers along with specific physiologic changes in the
cardiovascular system. Liu et al. (2007b) observed an increase in end-diastolic FMD and end-systolic
FMD, but decreases in end-diastolic basal diameter and end-systolic basal diameter in diabetics upon
exposure to PMi0. The authors also observed positive associations with FMD and blood pressure in
diabetic individuals. An examination of biomarkers found mixed results, with Liao et al. (2005) observing
an increase in vWF; Liu et al. (2007b) observing an increase in TBARS, but not CRP or TNF-a; and
Dubowsky et al. (2006) observing an increase in CRP and WBC s. Overall, it is unclear how differences
in each of the aforementioned biomarkers contribute to the potential overall cardiovascular effect
observed in diabetic individuals; however, an increase in inflammation, oxidative stress, and acute phase
response may contribute to cardiovascular effects.
Further examination of the potential effect of underlying diabetes in individuals exposed to PM
through epidemiologic studies has found some evidence, which supports an increase in
December 2008
8-15
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
cardiovascular-related ED visits and hospital admissions. Multicity studies have found upwards of 75%
greater risk of hospitalization for cardiac diseases in individuals with diabetes upon exposure to PM10
(Zanobetti and Schwartz, 2002). Additional single-city analyses have also found an increase in risk for
ED visits for IHD, arrhythmias, and CHF in response to PMi0 exposure in diabetics (Metzger et al., 2007;
Peel et al., 2007). However, some studies (both multicity and single-city) have not observed an
association between cardiovascular ED visits and hospital admissions in response to exposure to PMi0 in
diabetics (Pope et al., 2006; Wellenius et al., 2006b; Zanobetti and Schwartz, 2005). Overall, although
studies have reported health effects associated with PM in a study population of diabetics, further work is
needed to confirm these associations, and to investigate by which pathophysiological pathway(s) diabetics
may have a greater response to PM.
Table 8-2. Percent of the U.S. population inflicted with respiratory diseases, cardiovascular
diseases, and diabetes.
Age Regional
Adults (18+)* 18-44 45-64 65-74 75+ NE MW S W
Chronic Condition/ Number (x106) % % % % % % % % %
Disease
RESPIRATORY DISEASES
Asthma* 24.2 11.0 11.5 10.5 11.7 9.3 11.7 11.5 10.5 10.8
Asthma (< 18 years) 6.8* 9.3* — — ______
COPD
Chronic bronchitis 9.5 4.3 2.9 5.5 5.6 6.7 3.8 4.4 4.9 3.5
Emphysema 4.1 1.8 0.3 2.4 5.0 6.4 1.4 2.3 1.9 1.6
CARDIOVASCULAR DISEASES
All heart disease 24.1 10.9 3.6 12.3 26.1 36.3 10.8 12.7 10.9 9.2
Coronary heart disease 14.1 6.4 0.9 7.2 18.4 25.5 6.4 7.6 6.6 4.7
Hypertension 51.6 23.4 7.7 32.4 52.7 53.5 22.2 23.7 25.3 20.6
Stroke 5.6 2.6 0.5 2.4 7.6 11.2 2.1 2.8 2.9 2.2
Diabetes 17.1 7.8 2.6 10.4 18.2 17.9 7.2 8.1 8.0 7.4
* All data for adults except asthma prevalence data for children under 18 years of age, from CDC (2008a, b). For adults prevalence data based off adults responding to "ever told had
asthma."
Source: Pleis and Lethbridge-Qejku (2007); CDC (2008a, b)
In addition to diabetes, obesity has been examined as a health condition, which could potentially
lead to an increase in PM-related health effects. Although Schwartz et al. (2005a) found an adverse
modification of HRV in obese individuals and Dubowsky et al. (2006) observed an increase in
December 2008
8-16
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
inflammatory markers (i.e., CRP, IL-6, and WBC) in response to short-term exposure to PM2 5 it remains
unclear if obese individuals are more susceptible to PM-related health effects.
8.2.2. Vulnerability Characteristics
Epidemiologic studies have examined characteristics that potentially increase the vulnerability of
subpopulations to PM-related health effects by analyzing potential effect modification of the association
between health outcomes and PM exposure. In most cases those individuals vulnerable to PM-related
health effects do not have underlying conditions that result in increased susceptibility to PM exposure, but
instead are disproportionately exposed to PM.
8.2.3. Urban Environment
Zeka et al. (2005) in their analysis of 20 U.S. cities found that exposure to PM10 increased in
mortality as population density increased and as the percent of primary PM10 from traffic increased across
cities. Both of these factors taken together represent the urban environment, and support the hypothesis
that as density of the urban environment increases so does the percent of the primary PM10 from traffic
sources (Zeka et al., 2005). Likewise, Wilson et al. (2007b) found that the most urban environment had
the highest mortality risk estimate when stratifying Phoenix into central, middle, and outer rings of
varying urban density. Although urban density and traffic-related PM contribute to the vulnerability of
individuals that reside in urban environments, other factors (e.g., socioecominic status (SES) and
education level) are known to highly influence the overall health of a population and may also be
contributing to the effects associated with an urban environment.
8.2.4. Socioeconomic Status
Based on data from the 2000 U.S. census, from among commonly-used indicators of SES, about
12% of individuals and 9% of families are below the poverty level, and 20% of the U.S. population does
not have a high school or higher level of education (U.S. Census). These individuals classified as having a
lower SES and/or education level than the general population tend to reside in low-income areas, which
are sometimes located in urban environments. Laurent et al. (2008) and O'Neill et al. (2003) noted that
there has not been a consistently-observed trend that characterizes the impact of SES on exposure to PM
(sometimes including BC or sulfate aerosols) or other air pollutants. Laurent et al. (2008) argued that the
December 2008
8-17
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
spatial scale of the study and the number and nature of selected SES variables and within-variable factors
clearly impacts their influence in epidemiologic models (see Section 3.7.4).
SES and education level have been shown in some studies (e.g., Filleul et al., 2004; Finkelstein et
al., 2003) to modify health outcomes of PM exposure for a population and may also be contributing to the
effects associated with an urban environment. For instance, Franklin et al. (2008) noted an increased risk
in mortality upon exposure to PM2.5 and its components for individuals of low SES. Additional analyses
stratified by education level have also observed consistent trends of increased risk in mortality for all
PM-size fractions for individuals with low educational attainment (Ostro et al., 2006; 2008; Zeka et al.,
2006). However, other factors associated with low SES and educational attainment that were not
examined in any of the aforementioned studies, which may increase these individuals sensitivity to
PM-related health effects include: higher prevalence of pre-existing diseases; inadequate medical
treatment; and limited access to fresh foods leading to a reduced intake of antioxidant polyunsaturated
fatty acids and vitamins (Kan et al., 2008). It should be noted that in some other studies (e.g., Filleul et
al., 2005; Tolbert et al., 2000), no effect of SES on health outcome was demonstrated.
Air conditioning use, although not a direct measure of SES, has been found to reduce personal
exposure to PM. Barn et al. (2008) and Baxter et al. (2007b). Both noted that window opening was an
important variable for determining PM infiltration factors. Franklin et al. (2007), in an examination of
27 cities, observed a marked decrease in mortality as the percentage of homes using air conditioning
increased, with "the effect of PM2 5 disappearing] in communities with high central air conditioning
prevalence and summer peaking particle concentrations." However, the results from epidemiologic
studies have not overwhelmingly concluded that air conditioning use significantly reduces the
PM-mortality association. Zeka et al. (2005) in a multicity analysis only observed a slight decline in
PMio-mortality association as the percentage of the population with air conditioning increased. A more
recent analysis by Franklin et al. (2008) did not analyze air conditioning use individually because they
believed using air conditioning alone in a model did not adequately account for the differences in building
ventilation rates, which differ by season and community. Therefore, by using temperature as a surrogate
of ventilation, Franklin et al. (2008) found that at times when temperatures are extremely low or high the
PM2 5 mortality association was reduced potentially because of a reduction in building ventilation rates.
Although consistency has not been observed uniformly across studies regarding the effect of air
conditioning use on PM-related health effects, the evidence suggests that individuals with access to air
conditioning, and more than likely with a higher SES, could potentially be less vulnerable to PM-related
health effects.
December 2008
8-18
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
8.2.5. Geographic Location
In addition to the previously mentioned, economic and societal factors that could potentially
increase the vulnerability of an individual to PM-related health effects, numerous studies have also
continued to examine the geographic heterogeneity in PM-related health effects. All of the studies
identified in the current PM ISA that have examined the PM-mortality relationship, in regards to
geographic location within the U.S., have concluded that the effects are greater in the East compared to
the West. Although the definition of East versus West varies from study to study the effects observed are
fairly consistent regardless of the PM-size fraction analyzed. Dominici et al. (2007b) and Peng et al.
(2005) both found that the East exhibited a larger percent increase in all-cause (non-accidental) mortality
compared to the West when examining PMio-mortality effects. However, in addition to an all season
analysis, Peng et al. (2005) conducted seasonal analyses and observed the same pattern of effects, with
the greatest effects occurring in the Northeast during the summer along with some indication of a summer
effect in the industrial Midwest. In the examination of the PM2 5 -mortality relationship by geographic
location Franklin et al. (2007) and Franklin et al. (2008) found results consistent with those reported in
the PM10 analyses. Franklin et al. (2007) observed greater effects in the East compared to the West;
whereas, Franklin et al. (2008) found similar results, but compared effects observed in the East and
Central U.S. to those in the West. Overall, the epidemiologic literature suggests that individuals residing
in the Eastern U.S. are more vulnerable to PM-related health effects.
December 2008
8-19
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Chapter 9. Ecosystem and Welfare Effects
9.1. Introduction
This chapter is a concise synthesis and evaluation of the most policy-relevant science used to help
form the scientific foundation for review of the secondary (welfare-based) NAAQS aimed at protecting
against welfare effects of ambient airborne PM. Specifically, Chapter 9 assesses the effects of
atmospheric PM on the environment, including: (a) effects on visibility, (b) effects on individual
organisms, (c) direct and indirect effects on ecosystem, (d) effects on materials, and (e) effects on climate.
As discussed in chapter 1, the effects of particulate NOx and SOx have recently been evaluated in the ISA
for Oxides of Nitrogen and Sulfur - Ecological Criteria (U.S. EPA, 2008e). That ISA focused on the
effects of deposition of NOx and SOx (both gas- and particle-phase) related to acidifying deposition and
nutrient enrichment, as well as the potential for increased mercury methylation. Thus, in conjunction with
the ISA for NOx and SOx, emphasis here is placed on the effects of airborne PM on visibility and climate,
and on the effects of deposition of PM constituents other than NOx and SOx, primarily metals and
carbonaceous compounds.
Section 9.2 of this chapter provides the summary and conclusions for the major welfare and
ecological effects evaluated. EPA's framework for causality, described in Chapter 1, is applied throughout
the evaluation and the causal conclusions are highlighted in this first section. The evidence for each of
these major effects is evaluated in more detail in Sections 9.3 through 9.8. These sections initially
highlight the conclusions from the 2004 PM AQCD (U.S. EPA, 2004), followed by an evaluation of
recent publications and assessment of the expanded body of evidence. In some sections, few new
publications are available, and the discussion is primarily a brief overview of the key conclusions from
the previous review.
9.2. Summary and Conclusions
9.2.1. Summary of Effects on Visibility
The evidence is sufficient to infer a causal relationship between ambient PM and visibility
impairment. Visibility impairment is caused by light scattering and absorption by suspended particles and
gases. N02 is the only commonly occurring atmospheric pollutant gas that absorbs visible spectrum
December 2008
9-1
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
radiation, though in most situations the amount of light absorption by N02 is overwhelmed by the higher
levels of particulate light extinction (i.e. the combination of scattering and absorption) usually
accompanying high N02 concentrations. Light scattering by gases in a pollutant-free atmosphere provides
a limit to visibility in pristine conditions and is a major contributor to the total light extinction during the
least visibility-impaired periods in remote regions of the western U.S. PM is the overwhelming source of
visibility impairment. EC and some crustal minerals are the only commonly occurring airborne particle
components that absorb light. All particles scatter light, and generally light scattering is the largest of the
four light extinction components. While larger particles scatter more light than similarly shaped smaller
particles of the same composition, the light scattered per unit of mass concentration is greatest for
particles with diameters from about 0.3 to 1.0 |im.
For special studies where detailed particle composition by size data are available, accurate
calculations of light extinction can be made. However, routinely available PM speciation data can be used
to make reasonable estimates of light extinction using relatively simple algorithms that multiply the
concentrations of each of the major PM species by its dry extinction efficiency and by a water growth
term that accounts for particle size change as a function of relative humidity for hygroscopic species (i.e.
sulfate, nitrate and sea salt). This permits the visibility impairment associated with each of the major PM
components to be separately estimated from PM speciation monitoring data. There are six major PM
components: PM2 5 sulfate usually assumed to be ammonium sulfate, PM2 5 nitrate usually assumed to be
ammonium nitrate, PM2 5 organic carbon compound, PM2 5 EC, PM2 5 crustal material (call fine soil), and
PMiq-2 5 or coarse mass.
Particulate sulfate and nitrate are produced in the atmosphere from gaseous precursors, making
them secondary PM species. They both have comparable light extinction efficiencies (haze impacts per
unit mass concentration) at any relative humidity level, their light scattering per unit mass concentration
increases with increasing relative humidity, and at sufficiently high humidity levels (RH>85%) they are
the most efficient particulate species contributing to haze. Particulate sulfate is the dominate source of
regional haze in the eastern U.S. (>50% of the particulate light extinction) and an important contributor to
haze elsewhere in the country (>20% of particulate light extinction).
Particulate nitrate is a minor component of remote-area regional haze in the non-California western
and eastern U.S., but an important contributor in most of California and in the upper Midwestern U.S.
especially during winter when it is the dominant contributor to particulate light extinction. While both
nitric acid (a reaction product of NOx emissions) and ammonia are needed to form ammonium nitrate, the
apparent reason for the Midwest nitrate bulge is an abundance of atmospheric ammonia in this region
principally from agricultural emissions. There is evidence that transport from the Midwest nitrate bulge
region is responsible for some of the ammonium nitrate episodes experienced in downwind regions far to
December 2008
9-2
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
the east. Urban particulate nitrate concentrations are significantly elevated above surrounding remote-area
background concentrations with the largest urban contributions in the western U.S. Particulate ammonium
nitrate concentrations in California and the Midwestern nitrate bulge region are an order of magnitude
greater than estimated natural levels of ammonium nitrate. Thermodynamic and air quality simulation
modeling shows that particulate nitrate concentrations are sensitive to changes in either NOx emissions
(from a combination of mobile and point sources) and ammonia emissions (principally from agricultural
sources), with the responsiveness of particulate nitrate to emissions changes depending on the relative
abundance of ammonia and nitric acid in the atmosphere.
EC and organic compound species (i.e. carbonaceous components of PM) have the highest dry
extinction efficiencies of the major PM species and are responsible for a large fraction of the haze
especially in the Northwestern U.S., though absolute concentrations are as high in the eastern U.S. Both
are a product of incomplete combustion of fuels, including those used in internal combustion processes
(gasoline and diesel emissions) and open biomass burning (smoke from wild and prescribed fire). Organic
compound PM species are also produced by atmospheric transformation of precursor gaseous emissions.
Smoke plume impacts from large wildfires dominate many of the worst haze periods in the western U.S.
Carbonaceous PM is generally the largest component of urban excess PM2 5. western urban areas have
more than twice the average concentrations of carbonaceous PM than remote areas sites in the same
region. In eastern urban areas PM2 5 is dominated by about equal concentrations of carbonaceous and
sulfate components, though the usually high relative humidity in the East causes the hydrated sulfate
particles to be responsible for about twice as much of the urban haze as that caused by the carbonaceous
PM.
Radiocarbon dating of carbonaceous PM from twelve sites (eight in the West, two of which are
urban) showed that about half of the urban area carbonaceous PM is from contemporary as opposed to
fossil sources, while in remote areas the fraction that is contemporary ranges from 82%-100%. Summer
urban excess carbonaceous PM is dominated by fossil carbon for the two western urban areas (Phoenix,
AZ and Seattle, WA), but nearly half of the winter urban excess for these two urban areas are from
contemporary carbon sources (e.g. residential wood combustion). An empirical relationship between the
radiocarbon analysis results and the more widely measured elemental and organic carbon data set was
used to estimate the fraction of contemporary carbon at about 150 monitoring locations nationwide. The
highest fraction of contemporary carbon is for the western remote areas sites during the summer (>90%
contemporary) and the least was for eastern urban areas during the summer (< 45% contemporary).
Winter tended to have less extreme fractions of contemporary carbon for both remote and urban areas. A
lower bound estimate of 40% of the contemporary and 35% of the fossil carbon is from secondary
conversion of gaseous precursor during the summer at the twelve radiocarbon monitoring sites,
December 2008
9-3
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
suggesting that primary carbonaceous PM whether from fossil or contemporary sources represent less
than two thirds of the total carbonaceous PM.
PM2 5 crustal material (referred to as fine soil) and coarse mass (i.e., PM10 minus PM2 5) are
significant contributors to haze in the arid Southwestern U.S. where they contribute a quarter to a third of
the haze, with coarse mass usually contributing twice that of fine soil. Coarse mass concentrations are as
high in the Central Great Plains as in the Southwestern deserts though there are no corresponding high
concentrations of fine soil as in the Southwest. Also, the relative contribution to haze by the high coarse
mass in the Great Plains is much smaller because of the generally higher haze levels caused by the high
concentrations of sulfate and nitrate PM in that region.
A comprehensive assessment of the 610 worst haze sample periods over a three year period in the
western U.S. where dust is the major contributor categorized each site/sampler period into four causal
groups: Asian dust, local windblown dust, transported regional windblown dust, and undetermined dust
(i.e. not in one of the three other groups). Most dust days occurred at sites in Arizona, New Mexico,
Colorado, western Texas, and southern California, and these were dominated by local and regionally
transported wind-blown dust. Asian dust caused only a few of the worst dust days during the 3-year
assessment period, though it is an important source of dust for the more northerly regions of the West
(responsible for 10%-40% of their worst dust periods) were there is rarely any windblown dust probably
due to the greater ground cover. The frequency of worst dust events classified as undetermined was
greatest for sites in the vicinity of large urban and agricultural areas such as those in California and
Southern Arizona.
Urban visibility has been examined in two types of studies directly relevant to this review process:
urban visibility preference studies and urban visibility valuation studies. Both types of studies are
designed to evaluate individuals' desire for good VAQ where they live, using different metrics. Urban
visibility preference studies examine individuals' preference by investigating the basic question "what
level of visibility degradation is unacceptable," while economic studies examine preference by
investigating "how much would you be willing to pay to improve visibility."
There are three completed urban visibility research projects focused on identifying preferences for
urban VAQ, and one additional pilot study (designed as a survey instrument development project) that
provided additional information. The completed studies were conducted in Denver, Colorado (Ely et al.,
1993), two cities in British Columbia, Canada (1996), and Phoenix, Arizona
(BBC Research & Consulting, 2002). The pilot study was conducted in Washington, DC (ABT, 2001).
One notable finding of the three visibility preference studies and the one pilot study is the general degree
of consistency in the median preferences for an acceptable level of visibility degradation. The range of
median acceptable preference level from the four studies is 19 to 25 deciviews (DV), the preferred
December 2008
9-4
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
measure of visibility impairment. Measured in terms of visual range (VR), these median acceptable levels
are between 30 and 55 km.
The economic importance of urban visibility has been examined by a number of studies designed to
quantify the benefits (or willingness to pay) associated with potential improvements in urban visibility.
Urban visibility valuation research prior to 1997 was summarized in Chestnut and Dennis (1997), and
was also described in the 2004 PM AQCD (U.S. EPA, 2004) and the 2005 OAQPS PM NAAQS Staff
Paper (U.S. EPA, 2005b). Since the mid 1990s, little new information has become available regarding
urban visibility valuation.
9.2.2. Summary of Effects on Individual Organisms and Ecosystems
The evidence is sufficient to infer a causal relationship between ambient PM and effects on
individual organisms and ecosystems, based on information from the previous review and limited
new findings in this review. However, our ability to quantify relationships between ambient
concentrations of PM and ecosystem response is hampered by significant data gaps and uncertainties.
The PM deposition that can affect individual organisms and ecosystems is not a single pollutant.
Rather, PM deposition comprises a heterogeneous mixture of particles differing in origin, size, and
chemical composition. The effects of exposure to a given mass concentration of PM of a particular size
may, depending on the mix of deposited particles, lead to widely differing phytotoxic responses and
ecosystem effects.
The deposition of PM onto vegetation and soil, depending on its chemical composition, can
produce direct or indirect responses within an ecosystem. The ecosystem response to pollutant deposition
is a direct function of the level of sensitivity of the ecosystem and its ability to ameliorate resulting
change.
Upon entering the soil environment, PM pollutants can alter ecological processes of energy flow
and nutrient cycling, inhibit nutrient uptake, change ecosystem structure, and affect ecosystem
biodiversity. Many of the most important effects occur in the soil. The soil environment is one of the most
dynamic sites of biological interaction in nature. It is inhabited by microbial communities of bacteria,
fungi, and actinomycetes. These organisms are essential participants in the nutrient cycles that make
elements available for plant uptake. Changes in the soil environment that influence the role of the bacteria
and fungi in nutrient cycling determine plant and ultimately ecosystem response.
Changes in the soil can result from the deposition of heavy metals. Exposures to heavy metals are
highly variable, depending on whether deposition is by wet or dry processes. Few metals (e.g., Cu, Ni,
Zn) have been documented to cause direct phytoxicity under field conditions. Exposure to coarse particles
December 2008
9-5
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
of natural origin and elements such as Fe and Mg are more likely to occur via dry deposition, while fine
particles of atmospheric origin are more likely to contain elements such as Ca, Cr, Pb, N, and V.
Ecosystems immediately downwind of major emissions sources such as power generating, industrial, or
urban complexes can receive locally heavy deposition inputs. By negatively affecting litter
decomposition, heavy-metal accumulation can adversely influence nutrient cycling (U.S. EPA, 2004).
Phytochelatins produced by plants as a response to sublethal concentrations of heavy metals are
indicators of metal stress to plants. The increasing concentrations of phytochelatins across regions and at
greater elevation associated with gradients in levels of forest injury implicate them in forest decline
(U.S. EPA, 2004).
The amount of PM entering the immediate plant environment and deposited onto the plant surfaces
or soil in the vicinity of the roots, determines the biological effect. Three major routes are involved during
the wet and dry deposition processes: (1) precipitation scavenging in which particles are deposited in rain
and snow; (2) occult (fog, cloud water, and mist interception) deposition; and (3) dry deposition.
Deposition of PM on the surfaces of above-ground plant parts can have physical and/or chemical
effects. Particles transferred from the atmosphere to plant surfaces may cause direct effects if they reside
on the leaf, twig, or bark surface for an extended period; are taken up through the leaf surface; or are
removed from the plant via resuspension to the atmosphere, washing by rainfall, or litter-fall with
subsequent transfer to the soil.
An important characteristic of fine particles is their ability to affect the flux of solar radiation
passing through the atmosphere directly, by scattering and absorbing solar radiation, and, indirectly, by
acting as cloud condensation nuclei (CCN) that, in turn, influence the optical properties of clouds.
Regional haze has been estimated to diminish surface solar visible radiation by approximately 8%. Crop
yields can be sensitive to the amount of sunlight received, and crop losses have been attributed to
increased airborne particle levels in some areas of the world.
9.2.3. Summary of Effects on Materials
The evidence is sufficient to infer a causal relationship between ambient PM and effects on
materials. Building materials (metals, stones, cements, and paints) undergo natural weathering processes
from exposure to environmental elements (wind, moisture, temperature fluctuations, sunlight, etc.).
Metals form a protective film of oxidized metal (e.g., rust) that slows environmentally induced corrosion.
However, the natural process of metal corrosion is enhanced by exposure to anthropogenic pollutants.
A significant detrimental effect of particle pollution is the soiling of painted surfaces and other
building materials. Soiling changes the reflectance of opaque materials and reduces the transmission of
December 2008
9-6
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
light through transparent materials. Soiling is a degradation process that requires remediation by cleaning
or washing, and, depending on the soiled surface, repainting. Particulate deposition can result in increased
cleaning frequency of the exposed surface and may reduce the usefulness of the soiled material. Attempts
have been made to quantify the pollutant exposure levels at which materials damage and soiling have
been perceived. However, to date, insufficient data are available to advance our knowledge regarding
perception thresholds with respect to pollutant concentration, particle size, and chemical composition.
9.2.4. Summary of Effects on Climate
The evidence is sufficient to infer a causal relationship between ambient PM and effects on
climate, including both direct effects on radiative forcing and indirect effects that involve cloud
feedbacks that influence precipitation formation and cloud lifetimes. Direct effects are relatively
better understood than indirect effects. Aerosol PM can contribute to both atmospheric warming
(especially BC) and cooling (most other PM constituents); but the overall net effect is cooling, and this
partially counteracts the warming caused by greenhouse gasses. It also appears that PM can result in
precipitation suppression downwind of urban pollution sources. Orographic precipitation is
disproportionately affected, and this effect is believed to reduce orographic precipitation by up to 25% at
some locations in the western U.S.
9.3. Effects on Visibility
9.3.1. Introduction
In recent years, most visibility research involved characterizing visibility levels and trends,
improving our understanding of the atmospheric processes and pollutants responsible for the impacts, and
attribution of visibility-impairing pollutants to emission sources, source types, and regions. The
motivation for much of this work has come from the visibility protection provisions of the 1977 Clean Air
Act Amendments (CAAA) that called for the development of regulations to address reduction of regional
haze in 156 national parks and wilderness areas to natural levels, and from the subsequent Regional Haze
Rule (RHR) promulgated in 1999 by EPA in response to the CAA mandate. Implementation of the RHR
entails planned emissions reductions to reach natural haze levels by 2064 in six ten-year planning steps.
Haze levels caused solely by PM from natural sources are generally much lower than contemporary
levels, with the largest difference being between the inorganic salts ammonium sulfate and ammonium
December 2008
9-7
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
nitrate taken to be just a few tenths of a microgram per cubic meter each (Trijonis et al., 1990), while
current levels of both over large regions of the country are an order of magnitude or more larger (DeBell,
2006). However, natural source PM can be substantial on an episodic basis for crustal mineral PM
components during high windblown dust conditions and for carbonaceous PM from biogenic combustion
during wildfire and prescribed burning episodes. The need for information to generate RHR
implementation plans has resulted in extensive use of continental-scale air quality simulation modeling
and assessment of expanded ambient monitoring data sets.
Aside from the remote-area visibility investigations conducted in response to the RHR, relatively
little work on visibility effects has been done in recent years. However, the relationship between PM and
haze levels permits use of routine filter-based PM chemical speciation data collected in numerous urban
areas (Jayanty, 2003), as well as high time- and size-resolved PM speciation data available in several
cities such as those in the PM Supersites program (Solomon and Hopke, 2008) to improve our
understanding of urban visibility. Comparisons between urban and remote areas data in the same region
affords the opportunity to differentiate between regional and local visibility impacts. Better size and time
resolution PM composition data compared to that available from the routine monitoring programs reduce
the number of simplifying assumptions required to estimate visibility levels, thereby reducing the
uncertainty of the estimates.
There has been relatively little new research on the optical and human perceptual aspects of
atmospheric visibility over the last decade or more. These topics have been the subjects of numerous
earlier investigations that have been summarized in detail elsewhere (Middleton, 1952; Tombach and
McDonald, 2004; Trijonis et al., 1990; U.S. EPA, 1979, 1980; Watson et al., 2002), including past criteria
documents on PM, S02 and NOx (U.S. EPA, 1982, 1993, 2004).
The background section below contains an overview of long-available information to help provide
context to the more recently published literature summarized in subsequent sections.
9.3.2. Background
Air pollution-induced visibility impairment is caused by the loss of image-forming light (i.e.
signal) and the addition of non-imaging forming light (i.e. noise) between an observer and the object
being viewed. These changes to the light reaching the observer are a result of light being scattered and
absorbed by particles and gases in the sight path (see the schematic in Figure 9-1). Electromagnetic theory
developed to characterize the interaction of light with matter (Mie, 1908) permits the calculation of light
scattering and absorption by particles and gas molecules where the index of refraction and shape of
particles by size are known (van de Hulst, 1981).
December 2008
9-8
DRAFT-DO NOT QUOTE OR CITE
-------
1 The ability of human observers to visually detect distant objects or identify changes in their
2 appearance depends on the apparent contrast of the object against its background. The apparent contrast is
3 affected by changes in the transparency of the atmosphere caused by air pollution as well as factors not
4 related to air quality such as length of the sight path, scenic lighting and the physical characteristics of the
5 viewed object and other elements of the scene. To rigorously determine the perceived visual effects of
6 changes in the optical properties of the atmosphere requires the use of radiative transfer modeling to
7 determine changes in light from the field of view experienced by the observer, followed by the use of
8 psychophysical modeling to determine the response to the light by the eye-brain system. The complexity
9 of such an approach discourages its common use.
Characteristics of Observer
• Detection Thresholds
• Psychological Response to
Incoming Light
• Value Judgements
Light from clouds
scattered Into
sight path ^
Optical Characteristics of Illumination
• Sunlight (Sun Angle)
• Cloud Cover (Overcast. Puffy, etc.!
• Sky
Sunlight ^
scattered .. . . _T .
Light reflected
from ground
scattered into
sight path
Image-forming
light absorbed
Image-forming
light scattered /
out of sight path
Optical Characteristics of
Intervening Atmosphere
• Light Added to Sight Path by
Particles and Cases
# Image-Forming Light Subtracted
from Sight Path by Scattering
and Absorption
• Color
• Contrast Detail (Texture)
• Form
• Brightness
Source: Malm (1983)
Figure 9-1. Important factors involved in seeing a scenic vista are outlined. Image-forming
information from an object is reduced (scattered and absorbed) as it passes through
the atmosphere to the human observer. Air light is also added to the sight path by
scattering processes. Sunlight, light from clouds, and ground-reflected light all
impinge on and scatter from particulates located in the sight path. Some of this
scattered light remains in the sight path, and at times it can become so bright that the
image essentially disappears. A final important factor in seeing and appreciating a
scenic vista are the characteristics of the human observer.
December 2008
9-9
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Atmospheric light extinction is a fundamental atmospheric optics metric used to characterize air
pollution impacts on visibility. It is the fractional loss of intensity in a light beam per unit distance due to
scattering and absorption by the gases and particles in the air. Light extinction (bexl) can be expressed as
the sum of light scattering by particles (bSiP), scattering by gases (bSig), absorption by particles (ba,P) and
absorption by gases (bag). Light extinction and its components are expressed in units of inverse length,
typically either inverse kilometers (km1) or, as will be the convention in this document, inverse
megameters (Mm1). Traditionally, for visibility-protection applications, the most sensitive portion of the
spectrum for human vision (550 nm) has been used to characterize light extinction and its components.
A parametric analysis has shown that a constant fractional change in light extinction results in a
similar perceptual change regardless of baseline conditions (Pitchford et al., 1990). From this assessment,
the deciview haze index, which is a log transformation of light extinction, similar in many ways to the
decibel index for acoustic measurements, was developed (Pitchford and Malm, 1994). A one deciview
(ldv) change is about a 10% change in light extinction, which is a small change that is detectable for
sensitive viewing situations. The haze index in deciview units is an appropriate metric for expressing the
extent of haze changes where the perceptibility of the change is at issue. The regional haze rule has
adopted the deciview haze index as the metric for tracking long-term haze trends of visibility-protected
federal lands (U.S. EPA, 2001). Light extinction and it components are more useful metrics for
characterizing the apportionment of haze to its pollutant components due to the approximately linear
relationship between pollutant species concentrations and their contributions to light extinction.
Daytime visibility has dominated the perspective taken by those who have studied the visibility
effects of air pollution, though nighttime visibility is also know to be impacted by air pollution.
Stargazing is a popular human activity in urban and remote settings. The reduction in visibility of the
night sky is primarily dependent on the addition of a light into the sight path, the brightness of the night
sky, and the reduction in contrast of stars against the background (see the schematic in Figure 9-2). These
are controlled by the addition of PM, which enhances scattering, and the addition of anthropogenic light.
Scattering of anthropogenic light contributes to the "skyglow" within and over populated areas, adding to
the total sky brightness. The visual result is a reduction of the number of visible stars and the
disappearance of diffuse or subtle phenomena such as the Milky Way. The extinction of starlight is a
secondary and minor effect also caused by increased scattering. Anthropogenic light sources include
artificial outdoor lighting, which varies dramatically across space. Natural sources include the Moon,
planets, and stars that have a predictable rhythm across time.
December 2008
9-10
DRAFT-DO NOT QUOTE OR CITE
-------
starlight
absorbed
moonlight
scattering
by particles
starlight
scattered
urban tight polluti
particle
forward
scatter
particle
back
scatter
wildlife
affected by
unnaturally
bright
nighttime
sky
observer in a non-urban setting
reflectivo cloud
reflective cloud
starlight
absorbed
particle
side
scatter
moonlight p
scattering
by particles
starlight -
scattered I
bright urban nighttime sky
particle
forward
scatter
particle
back
scatter
Figure 9-2. Schematic of remote-area (top) and urban (bottom) nighttime sky visibility showing
the effects of PM and light pollution.
December 2008
9-11
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
The nighttime visual environment has some important differences to note. Light sources and
ambient conditions are typically five to seven orders of magnitude dimmer at night than in sunlight.
Moonlight, like sunlight, introduces light throughout an observer's sight path at a constant angle. On the
other hand, dim starlight emanates from all over the celestial hemisphere while artificial lights are
concentrated in cities and illuminate the atmosphere from below. Sight paths are often inclined upward at
night as targets may be nearby terrain features or celestial phenomena. Extinction behaves the same at
night as during the day, lowering the contrast of scenes through scattering and absorption; nevertheless
the different light sources will yield variable changes in visibility as compared to what has been
established for the daytime scenario. Little research has been conducted on nighttime visibility. Even if
the air quality-visibility interactions are shown to be similar between day and night settings, the human
psychophysical response at night is expected to differ. Recent advances in the ability to instrument and
quantify nighttime scenes (Duriscoe et al., 2007) can be utilized to evaluate and help establish standards
for nocturnal visibility. The remainder of this document focuses exclusively on daytime visibility.
9.3.2.1. Non-PM Visibility Effects
Light extinction due to the gaseous components of the atmosphere are relatively well understood
and well estimated for any atmospheric conditions. Absorption of visible light by gases in the atmosphere
is primarily by N02, and can be directly and accurately estimated from N02 concentrations by
multiplying by the absorption efficiency. Scattering by gases is described by the Rayleigh scattering
theory. Rayleigh scattering occurs in a pollution-free atmosphere as a result of light scattering by the gas
molecules that compose the atmosphere (i.e. N2, 02, C02, etc.) and depends on only on the density of the
atmosphere, with highest values at sea level (-12 Mm"1) and diminishing with elevation (8 Mm"1 at ~4
km), and varies somewhat at any elevation due to atmospheric temperature and pressure variations.
Rayleigh scattering can be accurately determined for any elevation and meteorological conditions.
N02 absorbs more light in the short wavelength blue portion of the spectrum than at longer
wavelengths. For this reason a plume or layer of N02 remove more of the blue light from the scene
viewed through the layer giving a yellow or brown appearance to the layer or plume. This filtering of blue
light by N02 can deepen the brown appearance of hazes over urban areas, although it is not the sole cause
of such discoloration (U.S. EPA, 1993). The photopic-weighted absorption efficiency at the 550 nm
wavelength is incorporated into the revised version of the algorithm for estimating light extinction from
aerosol data that is used for implementing the RHR (Pitchford et al., 2007). However, N02 is not
routinely measured at any of the monitoring sites representing visibility protected areas where its impacts
are assumed to be inconsequential compared to those of PM. At background concentrations N02
absorption is generally less than five percent of the light scattering by clean air (Rayleigh scattering),
December 2008
9-12
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
making it unperceivable. Plume visibility models are available to assess both achromatic and
discoloration associated with N02 light absorption, for point source emissions (Seigneur et al., 1984;
U.S. EPA, 1980).
9.3.2.2. PM Visibility Effects
Particle light extinction is more complex than that caused by gaseous components. PM is
responsible for most visibility impairment except under near-pristine conditions, where Rayleigh
scattering is the largest contributor to light extinction or in plumes of combustion sources that are
well-controlled for particulate emissions (e.g. coal-fired power plants with bag houses), where light
absorption by N02 may dominate the light extinction.
Light-absorbing carbon (e.g., diesel exhaust soot and smoke) and some crustal minerals are the
only commonly occurring airborne particle components that absorb light. All particles scatter light, and
generally particle light scattering is the largest of the four light extinction components. While a larger
particle scatters more light than a similar shaped smaller particle of the same composition, the light
scattered per unit of mass concentration [i.e. mass scattering efficiency in units of Mirf'/(|ig/m3) which
reduces to m2/g] is greatest for particles with diameters from about 0.3 to 1.0 |im. If the index of
refraction, particle shape and concentration as a function of particle size are well characterized, Mie
theory can be used to accurately calculate the light scattering and absorption by those particles. However,
it is rare that these particle properties are known, so assumptions are used in place of missing information
to develop a simplified calculation scheme that provides an estimate of the particle light extinction from
available data sets.
Particles composed of water soluble inorganic salts (i.e. ammoniated sulfate, ammonium nitrate,
sodium chloride, etc.) are hygroscopic in that they absorb water as a function of relative humidity to form
liquid solution droplet. Aside from the chemical consequences of this water growth, the droplets become
larger when relative humidity increases resulting in increased light scattering, hence the same PM dry
concentration produces greater haze levels. Figure 9-3 shows the effect of water growth as a function of
relative humidity on light scattering for two size distributions of ammonium nitrate and ammonium
sulfate particles as well as for internal and external mixtures (i.e. mixed within the same particle and in
separate particles respectively) of the two components. This figure illustrates a number of important
points. The water growth effect is substantial with an increase in light scattering by about a factor of ten
between 40% and 97% relative humidity for the same dry particle concentrations. The amount of
scattering is significantly dependent on the dry particle size distribution. However the growth curves for
ammonium sulfate, ammonium nitrate and mixtures of the two particle components are similar at any of
December 2008
9-13
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
the dry particle size distributions. Water growth curves are also available for sodium chloride, the major
component in sea salt, which is an important PM component at coastal locations.
0.1
CD
Q
(/J
X)
C
O
O
j-t—
it-
CD
O
O
O)
c
'i
£
cc
o
w
0.01
0.001
nh4no3
(NH4)2so4
External Mixture
(55.5% nitrate + 44.5% sulfate) c
o Internal mixture a/.-
(0.3 um, 1.
40
(0.6 um, 1.5)
50 60
70
% RH
80
90 100
Source: Tang (1996)
Figure 9-3. Effect of relative humidity on light scattering by mixtures of ammonium nitrate and
ammonium sulfate.
PM light scattering can be accurately calculated for any relative humidity if the chemical
composition as a function of dry particle size is known (Malm et aL 2007). However, most routinely
available ambient monitoring programs do not include data with sufficient detail to make such
calculations. The IMPROVE network with its greater than 150 remote area monitoring sites (DeBell,
2006) and the CSN (Jayanty, 2003) with its greater than 150 urban area monitoring sites collect 24-h
duration fine particle samples (PM2 5) that are analyzed for the major components including sulfate and
nitrate by ion chromatography. CSN also analyzes for ammonium ion, but does not monitor coarse mass
(PM ,0- while IMPROVE measures coarse mass but does not analyze for ammonium ion. Neither data
set has sufficient size resolution to make theoretical calculations of light extinction, nor does either
December 2008
9-14
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
program routinely monitor N02 concentrations, which would be required to calculate its contribution to
light extinction by absorption.
A simple algorithm is frequently used to estimate light extinction from the concentrations of the
major components. The concentration of each of the major aerosol components is multiplied by a dry
extinction efficiency value and for the hygroscopic components (e.g. ammoniated sulfate and ammonium
nitrate) an additional multiplicative term to account for the water growth to estimate that components
contribution to light extinction. Both the dry extinction efficiency and water growth terms are developed
by some combination of empirical assessment and theoretical calculation using typical particle size
distributions associated with each of the major aerosol components, and they are evaluated by comparing
the algorithm estimates of light extinction with coincident optical measurements. Summing the
contribution of each component gives the estimate of total light extinction. The most commonly used of
these is referred to as the IMPROVE algorithm because it was developed specifically to use the
IMPROVE aerosol monitoring data and was evaluated using IMPROVE optical measurements at the
subset of sites that make those measurements (Sisler et al., 1996). The formula for the traditional
IMPROVE algorithm is shown below.
Light extinction (bcxl) is in units of Mm"1, the mass concentrations of the components indicated in
brackets are in |ig/nr\ and f(RH) is the unitless water growth term that depends on relative humidity.
Since IMPROVE doesn't include ammonium ion monitoring, the assumption is made that all sulfate is
fully neutralized ammonium sulfate and all nitrate is assumed to be ammonium nitrate. Though often
reasonable, neither assumption is always true. In the eastern U.S. during the summer there is insufficient
ammonia in the atmosphere to neutralize the sulfate fully. Fine particle nitrates can include sodium or
calcium nitrate, which are the fine particle fraction of generally much coarser particles due to nitric acid
interactions with sea salt at near-coastal areas (sodium nitrate) or nitric acid interactions with calcium
carbonate in crustal aerosol (calcium nitrate) (Lee et al., 2008; Malm and Hand, 2007). Despite the
simplicity of the algorithm, it performs reasonably well and permits the contributions to light extinction
beXt ~ 3 X f(RH) x [Sulfate]
+ 3 x /(RH) x [Nitrate]
+ 4 x [Organic Mass]
+ 10 x [Elemental Carbon]
+ 0.6 x [Coarse Mass]
+ 10
(9-1)
December 2008
9-15
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
from each of the major components (including the water associated with the sulfate and nitrate
compounds) to be separately estimated.
The f(RH) term inflate the particulate sulfate and nitrate light scattering for high relative humidity
conditions. For relative humidity below 40% the f(RH) value is 1, but it increases to 2 at -66%, 3 at
-83%, 4 at -90%, 5 at -93% and 6 at -95% relative humidity. The result is that both particulate sulfate
and nitrate are more efficient per unit mass than any other aerosol component for relative humidity above
-85% where its total light extinction efficiency exceeds the 10m2/g associated with EC. Based on this
algorithm, particulate sulfate and nitrate are estimated to have comparable light extinction efficiencies
(i.e. the same dry extinction efficiency and f(RH) water growth terms), so on a per unit mass
concentration basis at any specific relative humidity they are treated as equally effective contributors to
visibility impacts.
9.3.3. Effects on Visibility
9.3.4. Monitoring and Assessment
Monitoring and the assessment of monitoring data serve a number of goals with regard to the
visibility effects of PM, including improving our understanding of the physio/chemical/optical properties
of the aerosol, characterizing spatial and temporal air quality patterns, and assessing the causes
(i.e. pollution sources and atmospheric processes) that are responsible for visibility impairment.
Information generated by special studies employing sophisticated instrumentation are typically needed to
advance our understanding of aerosol properties, while characterizing trends is the product of analyzing
routine monitoring data, whereas assessing the causes of haze usually involves a weight-of-evidence
approach applied to special study and/or routine monitoring data sets plus the use of air quality simulation
modeling. This section summarizes recently available information that is based on monitoring data.
9.3.4.1. Aerosol Properties
Particle size is the most influential physical property of aerosol with respect to their dry light
extinction efficiency. Chemical composition by size is used to ascertain density (needed to convert
aerodynamic to physical size) and to identify the water growth characteristics of the aerosol (needed to
calculate the particle size at ambient RH). To characterize aerosol properties of interest for visibility
effects, field monitoring programs typically include particle size distribution monitoring, high size
resolution particle sampling with subsequent compositional analysis, and optical monitoring. These
December 2008
9-16
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
generate data that permit optical closure assessments where the light scattering and/or light extinction
estimates from the aerosol data are compared to corresponding optical data. Since component
contributions to visibility are generally assessed by applying the IMPROVE or some similar algorithm to
measured or modeled aerosol concentration data, this section will include recent investigations that
evaluate or address various assumptions inherent in the use of these simple algorithms.
One component of the Big Bend Regional Aerosol and Visibility Observational (BRAVO) Study,
conducted at Big Bend National Park, TX in the summer and fall of 1999, entailed use of detailed
measurements of aerosol chemical composition, size distribution, water growth, and optical properties to
characterize the aerosol and assess the relationship between aerosol physical, chemical and optical
properties (Schichtel et al., 2004). Fine ammoniated sulfate during the BRAVO Study was about half the
fine particle mass concentration and was shown to be responsible for about 35% of the light extinction.
Rayleigh scattering was the second largest contributor at about 25%, followed by coarse particle (about
18%), and organic compounds (about 13%). There was little fine particle nitrate (less than 5% of the mass
concentration) and most of it is apparently in the form of sodium nitrate and two thirds of it was found in
the coarse mode where it comprises about 8% of the coarse particle mass concentration. Both the
composition of the nitrate and the fact of much of it being in the coarse size mode (2.5 |im >D>10 |im)
are inconsistent with the implied assumptions of the IMPROVE algorithm.
A year-long special study of coarse particle speciation was conducted at nine IMPROVE remote
area monitoring sites during 2003 to 2004 to provide additional information about the geographic and
seasonal variations in coarse particle composition (Malm et al., 2007). The same sampling and analytical
methodologies procedures were used for the PM10 samples as are routinely used on the IMPROVE PM2 5
samples. The IMPROVE coarse particle speciation study did not include ammonium analysis, so sulfate
and nitrate ions were assumed to be ammonium sulfate and ammonium nitrate. As expected crustal
minerals were the largest contributors to coarse mass overall (about 60%), though at Mt. Rainier the
fraction of coarse PM that was organic exceeded the crustal mineral by nearly two to one (i.e. 59.2%
compared to 33.5%) On average across sites the organic particulate contributed significantly at about one
quarter of the coarse mass, while ammonium nitrate was the third largest contributor to coarse mass
(about 8%). Seas salt was negligible overall, but high at the one coastal site (i.e. 12% at Brigantine, NJ).
The sites with the highest coarse nitrate concentrations are the two in California (San Gorgonio,
0.74 |ig/m3 and Sequoia, 0.69 |ig/nr') where fine nitrates are also high on average (2.66 |ig/m3 and
2.14 (ig/m3 respectively). Brigantine, a coastal site in New Jersey had the highest fraction of total nitrate
in the coarse size range (36%). The authors speculate that Brigantine's particulate nitrate is likely sodium
nitrate, the result of nitric acid reactions with sodium chloride. The nine-site average fraction of total
December 2008
9-17
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
nitrate in the coarse size range is 26%. By contrast, coarse sulfate concentrations are small with only
about -1% of the total sulfate in the coarse fraction.
Routine IMPROVE monitoring data include the mass concentration, but not the composition of the
coarse PM fraction, so the algorithm used to estimate light extinction doesn't include any provision for
varied coarse PM composition as shown in this study. This study shows that about 10% of the coarse
mass across the nine monitoring sites is composed of hygroscopic materials (i.e. ammonium sulfate,
ammonium nitrate and sea salt), which during high humidity conditions will scatter more light than
estimated by the current algorithm (e.g. -20% bias at -90% relative humidity). However, at coastal sites
such as the Brigantine, NJ, IMPROVE site where the combined concentration of the inorganic salts (i.e.
sea salt, nitrate and sulfate) constitute a significant fraction (-24% on average) of the coarse mass
concentration, the IMPROVE algorithm underestimation of light extinction by coarse PM can be
significant for high relative humidity conditions (-60% at -90% relative humidity). The resulting
underestimation of total light extinction is typically much smaller since fine particle light extinction
generally exceeds that contributed by coarse particles. Another issue with regard to estimating light
extinction from coarse PM concentration when the composition is not crustal minerals, as has been
assumed, has to do with the lower average density of the coarse mode particles that results in greater
particle numbers and/or larger particles and therefore a greater light extinction efficiency (Malm and
Hand, 2007).
Special studies with more complete, higher time resolution and size resolved particulate inorganic
ion species chemistry and precursor gases were conducted at seven of the nine sites with IMPROVE
coarse particle speciation monitoring (Lee et al., 2008). This work confirmed the presence of sodium and
calcium nitrate (referred to as mineral nitrate) primarily in the coarse particle size range in addition to fine
particle ammonium nitrate where low temperatures, high humidity and excess ammonium (beyond that
required to neutralize the particulate sulfate) favored particle phase equilibrium. Figure 9-4 is a map
showing the locations and sample times and estimated composition of the total particulate nitrate for the
seven locations for this special study. Sites with a high fraction of ammonium nitrate (e.g. San Gorgonio,
Bondville, and Brigantine) have the highest nitrate contributions to total mass concentration and haze
levels, whereas sites with high mineral nitrates tend to have low total nitrate contributions. This work
shows that the common assumption that particulate nitrate are in the fine particle size range and consists
principally of ammonium nitrate is not necessarily true.
December 2008
9-18
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Jiimdville (I etxiiaiy. 2003)
Yt&iRfriir« (liriy-'Aiuj 2002)
Brigantine (November, 2003)
¦GRSM (July/Aug. 2004),
San Gorgcrio (July. 2003)
~ NH,N0i
~ NaNOa
0 Ca(NOj)2
Source: Lee, et al (2008).
Figure 9-4. Estimated fractions of total particulate nitrate during each field campaign comprised
of ammonium nitrate, reacted sea salt nitrate (shown as NaNOs), and reacted soil dust
nitrate (shown as Ca(N03h)-
Extinction efficiencies for individual particle species can be theoretically calculated from
sized-resolved aerosol measurements and can be inferred using multiple linear regression applied to
aerosol composition and light extinction measurement data. In a recent publication, Hand and Malm
(2007) reviewed the literature since 1990 in which aerosol mass scattering efficiency values were
calculated or inferred. From these they have compiled normalized dry scattering efficiency values for the
individual species. Based on 93 separate determinations including marine, remote continental and urban
areas data sets, the average dry mass scattering efficiency for ammonium sulfate is 2.5 ± 0.6 m /g.
Average values tended to be somewhat lower for the marine aerosol (~2 nr/g) than for remote continental
(-2.7 nr/g) and urban (2.6 nr /g) areas, and values also tended to be lower for fairly clean arid locations
compared with more humid polluted areas.
Based on 48 separate determinations including remote area and urban area data sets, the average
dry mass scattering efficiency for ammonium nitrate is 2.7 ± 0.5 m2/g (Hand and Malm, 2007). Average
values were higher m remote locations (2.8 ± 0.5 m2/g) compared to urban locations (2.2 ± 0.5 m2/g)
though this might be accounted for by the predominate use of multiple linear regression for the remote
areas, which can be biased high, compared to the use of theoretical calculations for the urban data sets.
December 2008 9-19 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Organic fine PM extinction efficiency of 3.9+1.5 m2/g is based on 58 separate determinations,
though much higher values (~6 m2/g) resulted for locations influenced by industrial and biomass
combustion sources (Hand and Malm, 2007). These organic fine PM extinction efficiency values were
adjusted to use a consistent ratio of organic mass to organic carbon of 1.8 for each determination of the
mass concentration. This value is generally associated with aged organic PM, while for more freshly
emitted PM, such as in an urban environment, a smaller ratio (e.g. 1.4) would be more appropriate. This
could explain the discrepancy between two approaches used to estimate the organic PM light extinction
efficiency for Phoenix (Hand and Malm, 2006), which resulted in a significantly lower value where a site
specific regression method was used compared to the value obtained from a method optimized for
remote-area monitoring (2.47m2/g compared to 3.71 m2/g). However in Fresno both the mass balance and
light scattering balance was improved by using a ratio of 1.8 instead of 1.4 to estimate the organic
compound mass (Watson et al., 2007). Another possible or partial factor with respect to urban light
extinction efficiency for organic PM may be that the size distribution of freshly emitted organic PM in
urban areas extends significantly into the ultra-fine particle size range (Demeijian and Mohnen, 2008)
that is less efficient per mass concentration at light scattering than the generally larger-sized aged organic
PM such as from a distant forest fire as was measured at the Baltimore Supersite.
Hand and Malm (2007) also reviewed and made recommendation for extinction efficiencies for the
other components PM components including for mixed coarse mode (1.0 ± 0.9 m2/g based on 51
determinations) and fine mode dust or soil (3.3 ± 0.6 m2/g based on 23 determinations, but recommending
1.0 m2/g for use with data from realistic collection efficiency samplers) and fine sea salt (4.5 ± 0.9 m2/g
based on 25 determinations, but recommending 1.0 m2/g to 1.3 m2/g for use with data from realistic
collection efficiency samplers). This work did not address light absorption efficiency of elemental (or
black) carbon or crustal PM.
The Hand and Malm (2007) average dry mass light scattering efficiency values are generally
consistent with the values for the IMPROVE algorithm (as shown in equation 9-1). However the adoption
of the IMPROVE algorithm by EPA for calculating the haze metric used to track trends and assess the
nominal pace of progress for the Regional Haze Rule (U.S. EPA, 2001) resulted in much greater scrutiny
of its performance in estimating extinction (Lowenthal and Kumar, 2003; Malm and Hand, 2007; Ryan et
al., 2005). Among the issues raised is that the algorithm tended to underestimate the light extinction for
the haziest conditions that occur principally during the summer in the southeastern U.S. and overestimate
for near pristine conditions that tend to occur most often in the arid western U.S. Furthermore they
showed the lack of mass or light scattering closure at coastal sites due to sea salt that was not accounted
for by the IMPROVE algorithm. These assessments used mass concentration and light extinction closure
and regression analysis methods to infer that the dry extinction efficiency for the major fine particle
December 2008
9-20
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
components would need to vary in order to avoid the biased estimates of light extinction at the extremes.
Theoretical calculations of sulfate dry extinction efficiencies for 41 days of size-resolved chemical
composition data for Big Bend, Texas as part of the BRAVO Study produced a range of results from ~2.4
m2/g to ~4.1 m2/g, with the larger dry extinction efficiency values tending to be associated with higher
ammonium sulfate mass concentration and narrower size distributions (Schichtel et al., 2004).
In response to the technical concerns raised about the performance of the IMPROVE algorithm, a
revised algorithm was developed (Pitchford et al., 2007). The revised version of the algorithm differs
from the original algorithm by including a fine sea salt term related to the measured chloride ion
concentration, increases by about 30% the mass concentration of the organic aerosol component by
changing the ratio of organic compound mass to organic carbon mass from 1.4 to 1.8, uses site elevation
dependent Rayleigh scattering in place of lOMm"1 that had been used at every site, added a N02 light
absorption term and employs a split component model for the secondary particulate components (i.e.
sulfate, nitrate and organic species) with new water growth terms to better estimate their extinction at the
high and low extremes of the range. The revised algorithm is displayed below in equation 9-2 where bold
type face indicates terms that differ from the original IMPROVE algorithm (equation 9-1).
bext x 2.2 x fs(RH) x \Small Sulfate] + 4.8 x fL(RH) x [Large Sulfate]
+ 2.4 x fs(RH) x [Small Nitrate] + 5.1 x fL(RH) x [Large Nitrate]
+ 2.8 x [Small Organic Mass] + 6.1 x [Large Organic Mass]
+ 10 x [Elemental Carbon]
+ 1 x [Fine Soil]
+ 1.7 x fss{RH) x [Sea Salt]
+ 0.6 x [Coarse Mass]
+ Rayleigh Scattering {Site Specific)
+ 0.33 x [N02 {ppb)]
(9-2)
Small and large sulfate, nitrate and organic mass are used to refer to the splitting of the
concentrations of each of those three species into two size distributions. This approach accounts for
increased light extinction efficiency with mass by using a simple mixing model that assume that each of
these three components are comprised of an external mixture of small and large particle size modes.
Conceptually, the large mode particles represent aged or cloud-processed aerosol, while the small mode
particles represent relatively newly generated particles from the gas phase precursors. The former are
more likely to be associated with high concentrations while the latter are likely to be at relatively low
concentration.
December 2008
9-21
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
The geometric mean diameter and standard deviations assumed for these two size modes are
0.5 (mi and 1.5 for the large mode particles and 0.2 |im and 2.2 for the small mode particles. Mie theory
applied to these size distributions for the three species results in dry extinction efficiencies for the small
and large mode ammonium sulfate (2.2 m2/g and 4.8 m2/g), ammonium nitrate (2.4 m2/g and 5.1 m2/g)
and organic mass (2.8 m2/g and 6.1 m2/g). Water growth terms specifically derived for the small and large
size distribution using the upper branch of the hygroscopic growth curves for ammonium sulfate are
applied to both the sulfate and nitrate PM. No water growth is assumed for organic PM.
A simple empirically developed apportionment approach that was evaluated by testing the new
algorithms estimated light scattering at the 21 IMPROVE sites that have nephelometer-measured light
scattering data. The fraction of the fine particle component (sulfate, nitrate, or organic mass) that is in the
large mode is calculated by dividing the total concentration of the component by 20 (ig/m3 (e.g. if the total
fine particle nitrate concentration is 4 (ig/m3, the large mode concentration is 1/5 of 4 (ig/m3 or 0.8 (ig/m3,
leaving 3.2 (ig/m3 in the small mode). If the total concentration of a component exceeds 20 (ig/m3, all of it
is assumed to be in the large mode.
The performance of the original and revised IMPROVE algorithms was evaluated using the data
for 21 IMPROVE remote-area monitoring sites that also have nephelometer monitoring of particle light
scattering. Figures 9-5 and 9-6 are scatterplots of the estimated versus measured light scattering for the
two algorithms. The revised algorithm has noticeably reduced bias at the upper and lower extremes.
However, the new algorithm estimates have somewhat reduced precision (i.e., the points are more broadly
scattered). States have adopted the new algorithm for the technical assessments that support their
Regional Haze Rule State Implementation Plans, but the revised algorithm was too recently developed to
be incorporated into any of the peer-reviewed technical literature reported on below. In general the
differences resulting from use of the original versus the revised IMPROVE algorithm in identifying best
and worst haze conditions and the apportionment of the various PM components are small with exception
of coastal locations where sea salt may be a significant contributor.
December 2008 9-22 DRAFT-DO NOT QUOTE OR CITE
-------
0 50 100 150 200 250 300 350
Measured Bsp
Source: Pitchford, et al. (2007)
Figure 9-5. A scatter plot of the original IMPROVE algorithm estimated particle light scattering
versus measured particle light scattering.
350
300
0.
250
200
150
100
Q.
50
0
0
50
100
150
200
250
300
350
Measured Bsp
Source: Pitchford, et al. (2007).
Figure 9-6. Scatter plot of the revised algorithm estimates of light scattering versus measured
light scattering.
9.3.4.2. Spatial Patterns
1 The IMPROVE network is the basis for much of what is known about particulate species spatial
2 and temporal patterns for remote areas of the U.S. Though IMPROVE includes some urban monitoring
3 sites, most of what is known about urban particle speciation trends is based on the EPA Speciation Trend
4 Network (STN) and other similarly operated state particle speciation sites jointly referred to as the
5 Chemical Speciation Network (CSN) (Jayanty, 2003). The number of IMPROVE network sites has
6 increased considerably beginning in 2000, first to increase its ability to generate data representative of the
7 156 visibility-protected national parks and wilderness areas, then later as the states in the central U.S.
December 2008
9-23
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
Table 9-1. Regional Planning Organization websites with visibility characterization and source
attribution assessment information.
Type of
Information
Name and Web Address
RPO
Information Content and Comments
RPO Home Pages
Western Regional Air Partnership
http://www.wrapair.org/
WRAP
Organizational structure, plans, projects, reports and links to other
sites with additional information.
Central Regional Air Planning Association
http://www.cenrap.org/
CENRAP
MANE-VU works in close cooperation with Northeast States for
Coordinated Air Use Management (NESCAUM) and Mid-Atlantic
Regional Air Management Association (MARAMA) to develop the
Midwest Regional Planning Organization
http://64.27.125.175/mrpo.html
MRPO
technical information for Regional Haze Rule in the Northeast. All
three web sites contain unique technical support information.
Visibility Improvement State and Tribal
Association of the Southeast
http://www.vistas-sesarm.org/
VISTAS
Mid-Atlantic/Northeast Visibility Union
http ://www. manevu .org/
http://www.nescaum.org/topics/regional-haze
http://www.marama.org/visibility/
MANE-VU
NESCAUM
MARAMA
Visibility - Air Quality
Monitoring Data
Visibility Information Exchange Web Site
http://vista.cira.colostate.edu/views/
All RPOs
All IMPROVE and most other PM speciation data, RHR compatible
derived parameters, and user-friendly tools to summarize and
display data.
Emission Inventory
Data
Emissions Data Management System
http://www.wrapedms.org/default_login.asp
WRAP
WRAP emission inventory data warehouse and tools that provides a
consistent approach to regional emissions tracking
Monitoring Data
Assessment
Causes of Haze Assessment
http://www.coha.dri.edu/
WRAP
CENRAP
Monitoring site-specific descriptive characterizations and maps,
seasonal and trends analysis, air flow analysis, & receptor
modeling.
Visibility Modeling
U. of California-Riverside Modeling Center
http://pah.cert.ucr.edu/aqm/308/
http://pah.cert.ucr.edu/aqm/cenrap/index.shtml
http://pah.cert.ucr.edu/vistas/
WRAP
CENRAP
VISTAS
Descriptions of input data, performance, and results of regional
scale modeling (CMAQ & CAMx) & source attribution for base and
future year regional haze.
Integrated
Information to
Support RHR SIP
Preparations
Technical Support System
http://matar.cira.colostate.edu/tss/
WRAP
Provides access and common formats to display and summarize
emissions inventory information, monitoring data/ assessment and
regional haze modeling result to aid state and tribal analyst prepare
RHR implementation plans.
December 2008
9-24
DRAFT-DO NOT QUOTE OR CITE
requested additional remote-area monitoring to better understand their contributions to regional haze. The
expansion of the network into the central U.S. significantly improved our understanding of spatial trends
in a region of the country that had little speciation monitoring. Except as otherwise noted most of the
information in this section was from the IMPROVE Report IV (DeBell, 2006) and displays of data that
are readily generated using the Visibility Information Exchange Web Site (VIEWS). VIEWS, the ambient
monitoring data system, is one of several websites (as described in Table 9-1) sponsored by the Regional
Planning Organizations (RPO) that documents substantial, though often otherwise unpublished, technical
information generated to support implementation of the Regional Haze Rule.
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Figure 9-7 shows maps of remote area light extinction estimates from PM speciation data for two
years selected to demonstrate the additional information available due to the expansion of the IMPROVE
network into the central U.S. The locations of monitoring sites supplying the data shown as color contours
are shown as dot on the maps. Users of such contour maps are usually cautioned that the contours are only
there to guide the eye to sites with similar measurements and that nothing should be implied about spatial
patterns where there are no monitoring sites. Certainly these plots give proof to the wisdom of such
warnings. Prior to 2001 there were no IMPROVE or any other remote-area aerosol speciation monitoring
sites in the central states between northern Minnesota and Michigan to the north and Arkansas and
Kentucky to the south. The lack of monitoring over such a large region in the center of the country hid the
presence of high average regional haze levels over the Midwestern U.S. Smaller scale differences are seen
in the rest of the country and some of those are due to interannual variations as well as to better spatial
resolution made possible by a more dense monitoring network.
f'ACABl
2000 Annual
jaerosol_bext
2004 Annual
aerosol_bext
Source: VIEWS fhttp ://vi sta. ci ra. col ostate ,e du l\i iews/)
Figure 9-7. IMPROVE network PM species estimated light extinction for 2000 (left) and for 2004
(right).
Figure 9-8 shows the seasonal pattern of PM species estimated light extinction using maps of mean
values for each of the calendar quarter for 2004. The first quarter has the highest region of haze centered
in the Midwestern U.S.; the warmer second and third quarters have the region of highest haze over the
Ohio River Valley; and the fourth quarter is a composite with high haze in both the Midwest and Ohio
River Valley. Smaller regions of haze show up in the Columbia River Valley (border between Washington
and Oregon) in the colder first and fourth quarters and in Southern California in the warmer second and
third quarters.
December 2008
9-25
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
&CAD1
2004 1st Quarter
aerosol_bext
2004 2nd Quarter
aerosol_bext
foe AD 1
>ACAD1
2004 3rd Quarter
aerosol_bext
2004 4th Quarter
aerosol_bext
Source: VIEWS (http://vista.cira.coiostate.eduA/ iews/)
Figure 9-8. Mean estimated light extinction from PM speciation measurements for the first (top
left), second (top right), third (bottom left), and fourth (bottom right) calendar quarters
of 2004,
The IMPROVE algorithm permits each PM component contribution to light extinction to be
separately estimated. Figures 9-9, 9-10 and 9-11 display the seasonal variation of the percent contribution
to aerosol light extinction by the various component-estimates. Figure 9-9 shows the contributions by
sulfate and nitrate particulate including the haze enhancement caused by the absorbed water in humid
conditions. As shown in Figures 9-9, a large regional pattern of high contribution to haze by nitrate PM is
centered in the Midwest, and during the cooler months the nitrate PM is the dominant cause of haze in the
region responsible for a third to a half of the particulate light extinction. Midwestern particulate nitrate is
responsible for the regional pattern of the highest haze levels shifting from the Ohio River Valley during
summer to the Midwest in the winter as shown in Figure 9-8. Particulate nitrate is also a significant
contributor to particulate light extinction year-around in California, where is generally contributes
20%-40%. The Pacific Northwest, parts of Idaho and Utah experience large contributions to particulate
light extinction by nitrates during the colder seasons, with contributions of 20%-30%.
December 2008
9-26
DRAFT-DO NOT QUOTE OR CITE
-------
*ACAD1
WCAD1
2004 1 st Quarter
[F_NH4N03_bext
2004 1 st Quarter F
(NH4)2S04_bext
rACADl
2004 2nd Quarter
F_NH4N03_bext
2004 2nd Quarter
F_(IMH4)2S04_bext
ffl.C AD 1
2004 3rd Quarter
F_NH4N03_bext
2004 3rd Quarter
F_(NH4)2S04_bext
| ACAD1
rACADl
2004 4th Quarter
F_NH4N03_bext
12004 4th Quarter
I F_(NH4)2SQ4 bext
Source: VIEWS (http :IN\sta. ci ra. col ostate .e du A/ iews/)
Figure 9-9. Percent contributions of ammonium nitrate (left column) and ammonium sulfate (right
column) to particulate light extinction for each calendar quarter of 2004 (first through
fourth quarter arranged from top to bottom). Note that the color scales are different
for each map.
December 2008
9-27
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Figure 9-9 also shows that particulate sulfate is the predominate contributor in the eastern U.S.,
where it contributes 40% or more on average and during the summer months up to three quarters of the
particulate light extinction over much of the East. In the western U.S. particulate sulfate generally
contribute 20%-50% of the particle light extinction. Regions of the lowest fractional contributions by
particulate sulfate and nitrate for any calendar quarter are generally in the western U.S., and as are shown
in the subsequent two figures have significant contributions by crustal PM components (i.e. coarse mass
and fine soil) and by carbonaceous PM (i.e., organic mass and EC).
Figure 9-10 shows the contributions to haze by the carbonaceous PM components (i.e. organic
mass and EC). They show broadly similar patterns with the greatest contributions in the western U.S
especially during the warmer months of the year. For the most part this spatial pattern results from the
dominate contributions to haze by sulfate and nitrate PM in the eastern half of the U.S., leaving relatively
little for other component contributions. The fractional contribution to haze by organic PM is generally
two to five times that of EC. In absolute terms, both carbonaceous components tend to have two to three
times higher concentrations in the eastern U.S. than in the non-coastal western states.
Figure 9-11 shows the contributions to haze by coarse mass and fine soil components. As with the
carbonaceous components, these crustal dominated components have a similar spatial pattern with regions
of highest contribution to haze in the western U.S., and just as for the carbonaceous PM, the explanation
for low contributions in the eastern U.S. is the dominate contributions to haze by sulfate and nitrate PM
leaving relatively little for other components. The crustal components contribute more to haze in the arid
regions of the west including the southwestern deserts. In absolute terms, coarse mass concentrations are
as high in the rural areas of the center of the country (including Oklahoma, Arkansas, Kansas, Missouri,
and Iowa) as they are in the Desert Southwest. Typically coarse mass contributions to haze exceed those
of fine mass by a factor of 2 to 4.
December 2008
9-28
DRAFT-DO NOT QUOTE OR CITE
-------
2004 1st Quarter
F_OMC_bext
fftCADl
2004 2nd Quarter
F_OMC_bext
'acadi
2004 3rd Quarter
F_OMC_bext
rACADl
[2004 4th Quarter
F_QMC_bext
"acadi
2004 4th Quarter
FJECJaext
fACADl
2004 1 st Quarter
F_EC_bext
rACADl
2004 3rd Quarter
F_EC_bext
2004 1 st Quarter
F_EC_bext
'acadi
2004 2nd Quarter
F_EC_bext
Source: VIEWS (http :IN\sta. ci ra. col ostate .e du A/ iews/)
Figure 9-10. Percent contributions of organic mass (left column) and EC (right column) to
particulate light extinction for each calendar quarter of 2004 (first through fourth
quarter arranged from top to bottom). Note that the color scales are different for each
map.
December 2008
9-29
DRAFT-DO NOT QUOTE OR CITE
-------
"acadi
*AC AD 1
2004 3rd Quarter
F_CM_bext
fftCADl
[2004 4th Quarter
F_CM_bext_
2004 2nd Quarter I
,CM_bext |
•ACADI
2004 2nd Quarter
F_SOIL_bext
2004 1st Quarter
F_SOIL_bext
2004 3rd Quarter
F_SOIL_bext
"acadi
2004 4th Quarter
F_SOIL_bext
Source: VIEWS (http :IN\sta. ci ra. col ostate .e du A/ iews/)
Figure 9-11, Percent contributions of coarse mass (left column) and fine soil (right column) to
particulate light extinction for each calendar quarter of 2004 (first through fourth
quarter arranged from top to bottom). Note that the color scales are different for each
map.
December 2008 9-30 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
9.3.4.3. Urban and Regional Patterns
Using a combination of IMPROVE and CSN data, it is possible to compare urban PM2 <;
concentrations and composition to corresponding remote-area regional values. These are shown as paired
color contours maps for IMPROVE and IMPROVE plus CSN (see Figures 9-12 to 9-17) (U.S. EPA,
2004) used the pairing of IMPROVE and CSN monitoring sites at 13 selected urban areas to separate
local and regional contributions of three major PM2 5 components as shown in Figure 9-24. In Figures
9-12 and 9-13 we see that urban PMj< concentrations are systematically higher than those in the
surrounding non-urban regions. The urban excess is generally much higher in the western U.S. than in the
East (e.g. there are five contour intervals separating Salt Lake City from its remote regional area
compared to only two for Indianapolis). This implies that eastern and western urban PM2.5 concentrations
and resulting visibility levels are less different than the eastern and western regional concentrations and
visibility levels.
• IMPROVE Site
Hawaii
9.68
Alaska
¦ IMPROVE Urban Site
Puerto Rico /
Virgin Islands
8 47
7.26
6.05
4.84
3.63
12.42
1.21
0.00
(jg/m3
Source: DeBell et al. (2006).
Figure 9-12. IMPROVE Mean PM2.5 mass concentration determined by summing the major
components for the 2000 through 2004.
December 2008
9-31
DRAFT-DO NOT QUOTE OR CITE
-------
IMPROVE Site
IMPROVE Urban Site
Puerto Rico /
Virgin Islands
Alaska
Hawaii
30 7
¦ 14.7
13,0
11 4
9.78
8.15
6,52
4 89
13.26
1.63
0.00
(jg/m3
Source: DeBell et ai. (2006).
Figure 9-13. IMPROVE and CSN (STN) mean PM2.5 mass concentration determined by summing the
major components for 2090 through 2004
• IMPROVE Site
¦ IMPROVE Urban Site
Puerto Rico /
Virgin Islands
Source: DeBell et al. (DeBell, 2006).
Figure 9-14. IMPROVE mean ammonium nitrate concentrations for 2000 through 2004.
December 2008
9-32
DRAFT-DO NOT QUOTE OR CITE
-------
IMPROVE Site
IMPROVE Urban Site
o e
Puerto Rico /
Virgin Islands
Alaska
Hawaii
Source: DeBell et al. (2006).
Figure 9-15. IMPROVE and CSN (STN) mean ammonium nitrate concentrations for 2000 through
2004.
_T--
« _ ^ ^ 1 ^'
¦
6.63
5.46
4.86
^Esj [H / * L ' /¦¦¦: EBv
4.25
3.64
1
3.04
2.43
1.82
• xi y
^ICJU / K"1¦ ' .•••" «.{
1
1.21
0.61
0.00
1 —i
pg/m3
J? 1 'v-—l • IMPROVE Site
a- j: | IhJH ¦ IMPROVE Urban Site
O
1 Puerto
g#|Alaska Hawaii | Virgin Islands
Source: DeBell et al. (2006).
Figure 9-16. IMPROVE mean ammonium sulfate concentrations for 2000 through 2004.
December 2008
9-33
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
® Alaska
Hawaii
A STN Site
• IMPROVE Site
¦ IMPROVE Urban Site
© O
Puerto Rico /
Virgin Islands
Source: DeBell, et al. (2006).
Figure 9-17. IMPROVE and CSN (STN) mean ammonium sulfate concentrations for 2000 through
2004.
Figures 9-14, 9-15, and 9-24 show the PM2 5 nitrate in remote and urban areas. Here the western
states have urban particulate nitrate concentrations that far exceed twice the remote area regional
concentrations. For the Central Valley of California and Los Angeles areas, the urban excess of
ammonium nitrate exceeds regional concentrations by from 2 (.ig/m ' to 12 (ig/nr\ In the region of the
Midwest nitrate bulge, the urban concentrations were less than twice the regional concentrations for an
annual urban excess of about 1 ug/nr". Northeast and southeast of the Midwest nitrate bulge, annual urban
particulate nitrate concentrations are several tenths to about one microgram per cubic meter above the
remote area regional concentrations, with warmer southern locations tending to have the smaller
concentrations of both regional and urban excess particulate nitrate.
As shown in Figures 9-16, 9-17, and 9-24, annual-averaged urban particulate sulfate concentrations
are generally not much higher than the regional values, with urban excess generally of less than about a
half microgram per cubic meter. The exceptions apparent by comparing Figures 9-16 and 9-17 are in
Texas and Louisiana where urban excess particulate sulfate are greater than 1 ng/m". perhaps caused by
local emissions (e.g. from oil refineries). Urban contributions are a larger fraction of the total particulate
sulfate concentrations in the western U.S. because the regional levels are much lower than in the East.
The modest additional particulate sulfate concentrations associated with urban areas suggests that most
particulate sulfate is regionally distributed, and that IMPROVE and CSN monitoring sites can be used
December 2008
9-34
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
together to enhance our ability to delineate particulate sulfate spatial distributions. For example, note that
the additional data from urban sites shown in Figure 9-17 extends to the north and south the apparent
distribution of the high particulate sulfate loading shown in Figure 9-16 over Tennessee and Kentucky, as
well as the high loadings over southern Pennsylvania, eastern West Virginia and northern Virginia. (The
color-contour suggested dip in concentrations between the two eastern particulate sulfate high
concentrations regions may not exist in the atmosphere, but this cannot be verified without speciation
monitoring sites in southern Ohio, the boarder of Kentucky and West Virginia and western Virginia.)
Urban and remote area carbonaceous PM2.5 are displayed in Figures 9-18 and 9-19 (organic mass),
9-20 and 9-21 EC, and 9-22 and 9-23 (total carbon = organic + EC concentration). Just as with particulate
nitrate both organic mass and EC concentrations are more than twice the remote-area background
concentrations for western urban monitoring locations. One of the more interesting pairing of sites is for
the Virgin Islands compared to the urban site at San Juan Puerto Rico (see the map cutout Figures 9-18
through 9-21). The San Juan urban excess organic carbon is moderate, while the EC value is among the
most extreme inferred in this manner. For eastern urban areas, approximately half the total carbon is local
with the other half is regional. In eastern urban areas, carbonaceous and sulfate particulate are the two
major components of PM25, with roughly equal contributions, and account for over 80% of the mass
concentration. Edgerton et al. (2004) showed that carbonaceous PM2 5 is responsible for most of the urban
excess above regional concentrations at four urban/rural paired SEARCH monitoring sites in the
southeastern U.S. However, the higher overall light extinction efficiency for sulfate resulting from its
hydrophilicity gives it ~ 2: 1 dominance in responsibility for eastern urban light extinction.
Urban and remote area fine soil PM2 5 concentrations are displayed in Figures 9-22 and 9-23. Urban
fine soil levels are at most a few tenths of a microgram per cubic meter higher than the regional
background levels and in some regions they are much less. Just as with carbonaceous PM2 5, the Virgin
Island, San Juan, Puerto Rico pair are interesting for fine soil. In this case the interesting feature is that
both of these island monitoring sites have high concentrations of fine soil, which is caused to their being
in the trans-Atlantic transport path of dust from Africa (Prospero, 1996).
No urban - remote pair of coarse mass concentration maps is available because CSN does not
monitor coarse mass. Malm et al. (2004) contains a map of annual mean coarse mass concentration for
2003 which includes the values for IMPROVE urban sites, including two in the western U.S. with much
more coarse mass than the nearby remote areas monitoring sites (i.e., -24 |ig/m3 compared to ~9 |ig/m3
for Phoenix, and ~6 |ig/m3 compared to ~2 |ig/m3 for Puget Sound) and one eastern IMPROVE site at
Washington, DC with less coarse mass than the surrounding remote area values (~2 (.ig/nr3 compared to
~4 |ig/m3).
December 2008
9-35
DRAFT-DO NOT QUOTE OR CITE
-------
• IMPROVE Site
¦ IMPROVE Urban Site
Puerto Rico /
Virgin Islands
Alaska
Hawaii
i ¦ 5.79
¦3.72
3.30
2.89
2.48
2.07
1.65
1.24
10.83
0.41
0.00
(jg/rrv
Source: DeBell et al. (2006).
Figure 9-18. IMPROVE monitored mean organic mass concentrations for 2000 through 2004.
m 12.4
¦6.47
5.75
5 03
4.31
3.60
2.88
2.16
11.44
0.72
0.00
|jg/m3
• IMPROVE Site
¦ IMPROVE Urban Site
O O
Puerto Rico /
Virgin Islands
Alaska
Hawaii
Source: DeBell et al. (2006).
Figure 9-19. IMPROVE and CSN (STN) mean organic mass concentrations for 2000 through 2004.
December 2008
9-36
DRAFT-DO NOT QUOTE OR CITE
-------
• IMPROVE Site
Puerto Rico/
Virgin Islands
Alaska
Hawaii
IMPROVE Urban Site
_ 0.96
0.44
0.39
0.34
0.29
0.25
0.20
0.15
10.10
0.05
0.00
pg/m3
Source: DeBell et al. (2006).
Figure 9-20. IMPROVE mean EC concentrations for 2000 through 2004.
Alaska
Hawaii
IMPROVE Site
Urban IMPROVE Site
O O
Puerto Rico /
Virgin Islands
_ 1.74
-! 0 95
0.84
0.74
0.63
0.53
0.42
0.32
10.21
0 11
0.00
pg/rn3
Source: DeBell et al. (2006).
Figure 9-21. IMPROVE and CSN (STN) mean EC concentrations for 2000 through 2004.
December 2008
9-37
DRAFT-DO NOT QUOTE OR CITE
-------
• IMPROVE Site
Hawaii
¦ IMPROVE Urban Site
Puerto Rico/
Virgin Islands
¦ 3.08
1.44
1.28
1.12
0.96
0.80
0.64
0.48
10.32
0.16
0.00
pg/m3
Source: DeBell et al. (2006).
Figure 9-22. IMPROVE mean fine soil concentrations for 2000 through 2004.
A STN Site
Alaska
Hawaii
IMPROVE Site
IMPROVE Urban Site
© ©
Puerto Rico/
Virgin Islands
4 80
1 1.29
1.14
1.00
0.86
0.72
0.57
0.43
10.29
0.14
0.00
pg/m3
Source: DeBell et al. (2006).
Figure 9-23. IMPROVE and CSN (STN) fine soil concentrations, 2000 through 2004.
December 2008
9-38
DRAFT-DO NOT QUOTE OR CITE
-------
Nitrates
Sulfates
Fresno
Missoula
Salt Lake City
i |
WEST
Tulsa
I I
EAST
St. Louis
¦I
Bi rmlngham
XQ
Indianapolis
Atlanta
Cleveland
Charlotte
Richmond
5]
~ Regional
Contribution
Baltimore
~ Local
Contribution
New Yor k City
0 2 4 6 8
Annual Average Concentration
of Nitrates, ug/m3
t r
10 12
Fresno
Missoula
Salt Lake City
Tulsa
St. Louis
Birmingham
Indianapolis
Atlanta
Cleveland
Charlotte
Richmond
Baltimore
New York City
Fresno
Missoula
Salt Lake City
Tulsa
St. Louis
Birmingham
Indianapolis
Atlanta
Cleveland
Charlotte
Richmond
Baltimore
New York City
Carbon
XI
WEST
EAST
~ Regional
Contribution
~ Local
Contribution
i—r
2
~i—r
4 6 8
Annual Average Concentration
of Sulfates, ug/m3
"1—r
10
12
WEST
I I
EAST
I I
|
|
| |
I
1 ~ Regional
Contribution
|
~ Local
—1 Contribution
I
2 4 6 8 10 12
Annual Average Concentration
of Carbon, jig/m3
Source: U.S. EPA (2004)
Figure 9-24. Regional and local contributions to annual average PM2.5 by particulate sulfate, nitrate
and total carbon (i.e. organic plus EC) for select urban areas based on paired
IMPROVE and CSN monitoring sites.
December 2008
9-39
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
• IMPROVE Site
Hawaii
¦ IMPROVE Urban Site
Puerto Rico /
Virgin Islands
¦ 22 1
8.92
7.93
6.94
5.95
4.96
3.96
2.97
11.98
0.99
0.00
(jg/m3
Source: DeBell et al. (2006).
Figure 9-25. IMPROVE mean coarse mass concentrations for 2090 through 2004.
Figure 9-25 shows the remote area coarse mass concentrations as measured by the IMPROVE
network. The pattern of high coarse mass concentrations from Oklahoma to Iowa is comparable to the
high concentration levels in the desert southwest, though as shown in Figure 9-11 it contributes a smaller
relative share of the light extinction because of the higher contributions to haze by particulate nitrate and
sulfate m this agricultural region of the country. Comparing Figures 9-22 and 9-25 shows that the coarse
mass and fine soil concentration patterns are similar for the desert southw est but there is a much lower
fine soil to coarse mass concentration ratio for the agricultural center of the country, suggesting a regional
difference in the size distribution of suspendable soil materials.
9.3.4.4, Temporal Trends
Visibility trend analysis requires relatively long data records to avoid having meteorologically
driven interannual variability obscure more meaningful emissions-driven air quality trends. A requirement
for long-term data limits the number of monitoring sites useful for trend analysis. Maps that show haze
trends for IMPROVE sites for the 10-year period from 1995 through 2004 for the mean of the 20% best
and the 20% worst haze days where sites are required to have a minimum of six complete years of data
during the ten-year period is shown in Figures 9-26 and 9-27, respectively. The best haze days have
improving haze at most sites (32 of 47), no trend at several sites (10 of 47) and degrading visibility at just
December 2008
9-40
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
one site (Great Sand Dimes, CO). The worst haze days have improving haze conditions at several sites
(13 of 47), no trend at most sites (30 of 47) and degrading visibility at a few western sites (4 of 47).
Eight- ten- and sixteen-year trends analysis conducted for the Western Regional Air Partnership
(WRAP) as part of the Causes of Haze Assessment (http://www.wrapair.or) show that improving trends
for the 20% best haze levels for the sites in the western U.S. generally correspond to improving trends for
all of the major components with the exception of particulate nitrate. Trends assessment for the worst
haze days at western sites show consistent reductions in particulate sulfate, but otherwise have mixed
increasing and decreasing haze component trends, many of which are not statistically significant.
Edgerton, et al. (2004) showed a decreasing trend in PM2 5 of about 18% (corresponding to 1 iig/m" to
2 (.ig/m ) for four urban - airal paired SEARCH sites in the Southeastern U.S. corresponding to similar
reductions in sulfate and carbonaceous particulate.
Srioqua mie Pi
¦ jl
Irigantine
Canyonl
Pii
myon ^
> re at Smoky Mtns
¦ San Gj
Cape Romain
Improving Trend, p<=0.05
Improving Trend, 0.05
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
¦ Crater Lake
Bridger
igantine
Rocky Mi
CanyonJ
Sequoia
¦at Smoky Mtns
Sarij Gorgon
Up ser Buffalo
Cape Romaii
lalupe Mtns
Big
Improving Trend, p<=0.05
Improving Trend, 0.05
-------
1
2
3
4
5
6
7
8
9
10
decreasing particulate sulfate concentration trends and a correspondence in trends between S02 emissions
and particulate sulfate concentration by region (Holland et al., 1999; U.S. EPA, 2004).
Western US
North Eastern US
16 2
13.5
0,92 in
W 0 75
(0 4.5
10.8
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
from particulate nitrate and gaseous nitric acid, but nitric acid does not contribute to light extinction. For
situations with limited atmospheric ammonia or elevated temperatures, trends in nitrogen may be
principally in nitric acid with no net change in nitrate light extinction. Alternately with abundant ammonia
and low temperatures the trend in nitrogen may be principally in particulate nitrate and the nitrate
component of haze.
Ten-year trends (1994-2003) of particulate nitrate contribution to light extinction during the 20%
worst haze conditions conducted as part of the Causes of Haze Assessment (see the link in Table 9-1) are
shown in Figure 9-29. This indicates that haze from particulate nitrates is increasing across the western
U.S. at a rate of several Mm"1 per year in parts of California and at a rate of several tenths of an Mm"1
across the Four-Corners states. While statistically significant, these trends are influenced by an
unexplained nationwide period of depressed nitrate concentrations measured by the IMPROVE network
during a four year period from the winter of 1996-97 through the winter of 2000-01. Extensive
examinations of plausible monitoring methodological explanations have failed to offer any evidence that
the data are invalid (McDade, 2006), but no satisfactory atmospheric or emissions-related explanation has
been offered to account for this four-year depression of nitrate. Similar analyses of particulate nitrate haze
trends are not available for the rest of the country.
Maps of remote-area 8-, 10-, and 16-year trends for carbonaceous and crustal PM species based on
IMPROVE monitoring are available for the western U.S. conducted as part of the Causes of Haze
Assessment. Generally these show a broad range of results (i.e., a mixture of statistically significant
upward or downward trends and insignificant trends often with neighboring sites having opposing trends)
that vary considerably depending on the number of years selected (i.e. 8, 10, or 16) and whether trends are
for the best, worst, or middle of the haze distribution data. The scatter in these results is undoubtedly due
to the high interannual variability and varying locations of wildfire and wind-suspended dust emissions
that dominate the remote-area concentrations of the carbonaceous and crustal PM species in the western
U.S.
December 2008
9-44 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
# *
> \ I 1 rtkr
h-r~~l ui* iSfTh rM§*
# *t j 11 vU-ri-'r'\f
ft A f[ ft Legend
ft "k A # WNSIope
if " ft • ^ 0064 •0093
ft . 0.094 - 0.170
> 4 »/ *
#
rr^^ry *
ft o«,.
JL-r-J t ) . t «
.—-A
0.171 -0.520
1 509
i >
# VJ ^
c^-" Hawaii Virgin Islands.
Source: Causes of Haze Assessment website.
Figure 9-29. Map of 10-year trends (1994-2003) in haze by particulate nitrate contribution to haze
for the worst 20% annual haze periods. The orientation, size and color of the arrows
indicate the direction, magnitude and statistical level of significance of the trends.
These consistent upward trends may be a misleading result due to an unexplained
sampling issue (see text for additional information).
9.3.4.5. Causes of Haze
As indicated above, estimates of haze levels contributed by particulate species are proportional to
their concentrations at any relative humidity. However, in order to attribute haze to emissions from
individual sources, source types, or source regions, scientists generally apply any of a number of receptor
and air quality simulation modeling approaches and when using multiple approaches they reconcile the
results of each using a weight-of-evidence methodology. Commonly this methodology has been applied to
the extensive datasets generated by special studies designed to estimate source-receptor relationships for a
few receptor locations or for individual emission sources (Pitchford et &L 1999; 2005; Schichtel et aL
2005). More recently the Regional Planning Organizations (RPOs) have sponsored extensive regional
haze source attribution assessments using weight-of-evidence methodologies to reconcile attribution
results for virtually all of the remote-area IMPROVE sites to support the development of State
Implementation Plans for the Regional Haze Rule. Additionally, a number of recent urban special studies,
December 2008
9-45
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
including those sponsored by EPA PM Supersites program (Solomon and Hopke, 2008), have addressed
the causes of and sources contributing to urban excess haze above region levels.
The relative importance of the PM species that contribute to haze varies by region of the U.S. and
time of year as seen in Figures 9-9, 9-10 and 9-11, above. Generally haze in the western half of the U.S. is
not dominated by any one or two PM species. In the eastern half of the U.S., nitrate, especially during
summer and winter in the Midwest, and sulfate are the dominate haze species. As described above, urban
haze can be viewed as a composite of the regional and local contributions where local contributions seem
to be dominated by carbonaceous and to a lesser extent nitrate and crustal PM components. There have
been far fewer urban investigations that explicitly consider visibility impacts, though there are numerous
studies of urban PM source attribution. The order of discussion below on the cause of haze is by region
beginning in the western U.S. and proceeding to the east, analogous to dominate air flow patterns across
the lower 48 states and will include information from urban studies along side those of remote-area haze
investigations.
Based on modeling of an episode (Sept. 23 to 25, 1996) in the California South Coast Air Basin
(SCAB) and another episode (Jan. 4 to 6, 1996) in the Jan Joaquin Valley (SJV) by Ying and Kleeman
(2006), about 80% of the particulate sulfate for both regions are from upwind sources, with most of the
remaining associated with diesel and high-sulfur fuel combustion. Kleeman et al. (1999), using a
combination of measurements and modeling, showed that the upwind particulate sulfate source region for
the SCAB was over the Pacific Ocean (confirmed by measurements on Santa Catalina Island) and that
these particles subsequently grew with accumulation of additional secondary aerosol material, principally
ammonium nitrate as they traversed the SCAB. The majority of the nitric acid that forms particulate
nitrate in the SCAB is from diesel and gasoline combustion (-63%), while much of the ammonia is from
agricultural sources (-40%) and catalyst equipped gasoline combustion (-16%) and upwind sources
(-18%). The majority of the organic carbon found in SCAB was attributed in this study to primary
emissions by transportation-related sources including diesel (-13%) and gasoline (-44%) engines and
paved road dust (-12%). At the Fullerton site in the middle of the SCAB the concentration of locally
generated organics is roughly double that of the locally generated nitrates (-5.6 |ig/m3 compared to
-2.4 |ig/m3). while at Riverside on the east edge of the SCAB and near the large agricultural sources of
ammonia emissions, the particulate nitrate concentrations are nearly double that of organic PM
(-17 |ig/m3 compared to -10 |ig/m3).
Ying and Kleeman (2006) showed that during the winter 1996 episode in the SJV most of the nitric
acid that forms particulate nitrate is from upwind sources (-57%) with diesel and gasoline combustion
contributing most of the rest (30%), while much of the ammonia is from upwind sources (-39%) and a
combination of area, soil and fertilizer sources (-52%). In an assessment of PM particle size and
December 2008
9-46
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
composition in the SJV during the winter of 2000-2001, Herner, et al. (2006) showed that fresh emissions
of carbonaceous PM from combustion sources in urban locations (Sacramento, Modesto, and Bakersfield,
CA) move quickly from ultrafine particle size (i.e. diameter -0.1 |im) to accumulation mode by
condensation with accumulation mode (i.e., diameter -0.5 |im) particles, and that secondary nitrate
particle formation occurs preferentially on the surface of hydrated ammonium sulfate particles during the
afternoon when gas-phase nitric acid is at peak photo-chemical production from NOx. Given the
abundance of ammonia emissions and low ambient temperatures, particulate nitrate production in this
way is only limited by the availability of nitric acid. Due to the cool winter conditions there was little
secondary organic aerosol production during this study. Sea salt was shown to dominate the larger coarse
particle mode during on-shore wind at the background coastal monitoring site at Bodega Bay, north of
San Francisco, CA.
Using a regression analysis to find the dependence of particulate sulfate concentration measured
over a three year period (2000-2002) at 84 western IMPROVE monitoring sites on the modeled transport
trajectories to the sites for each sample period, Xu et al. (2006) were able to infer the source regions that
supplied particulate sulfate in the western U.S. Among the source regions included in this analysis were
the near coastal Pacific Ocean (i.e. a 300 km zone off the coast of California, Oregon, and Washington).
Up to half of the particulate sulfate measured at Southern California monitoring sites was associated with
this source region. As shown in Figure 9-30 the zone of impact from this source region included large
regions of California, Arizona, and Nevada. The authors made the case that high sulfur content fuel used
in marine shipping and port emissions may be largely responsible. As a result, the Western Regional Air
Partnership (WRAP) RPO emissions inventory was modified to include marine shipping and a Pacific
Offshore source region was added to source attribution by air quality simulation modeling.
The sulfate attribution results of the WRAP air quality modeling (results available on the Technical
Support System (TSS) website, see Table 9-1 for the web-link) credit the Pacific offshore source region
with somewhat smaller contributions than those from the trajectory regression work by Xu, with
concentrations at the peak impact site in California that are about 45% compared to 50% by regressions
and even greater differences for more distant monitoring sites. Based on the modeling attribution the
Pacific offshore source region was responsible for 10%-20% of the nitrate measured in Southern
California.
December 2008
9-47
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
m
-1! 1L
Contribution of the Pacific Coast to Sulfate Concentration
<0.1
0.1-0.2
¦ 0.2-0 4
¦ 0.4-0.6
¦ 0.6- 0.8
¦ 0.8- 1.4
Source: Xu et ai. (2006).
Figure 9-30. Contributions of the Pacific Coast area to the ammonium sulfate (pg/m3) at 84
remote-area monitoring sites in western U.S. based on trajectory regression (dots
denote locations of the IMPROVE aerosol monitoring sites).
A coordinated effort by federal, state, and county air quality organizations to determine the causes
of haze in the Columbia River Gorge (a deep and narrow gap in the Cascade Mountains on the
Washington/Oregon border) through extensive multiyear measurements and high spatial resolution air
quality modeling of typical episodes demonstrated the multitude of emission sources that contribute to its
impairment (Pitchford et al., 2007). During the summer, Gorge winds are generally from the west and
relatively dry. More than half of the haze during a typical summer episode is from a combination of
international and other distant sources (-22% at the western end of the Gorge) and regional natural
sources including wildfire and secondary (-39% at the western end of the Gorge) organic PM from
biogenic emissions. The Portland/Vancouver metropolitan area was responsible for a significant amount
of the haze during the summer (-20% on in the western end of the Gorge, while sources within the Gorge
were responsible for a moderate amount of haze (-6% and ^9% at the western and eastern ends of the
Gorge). The wind is much more often from the east during the winter. The highest haze levels in the
Gorge are dunng the winter and are associated with fog conditions that rapidly convert precursor gaseous
emissions of NOx and S02 from local and regional combustion sources and NH4 from local and regional
agricultural activities to secondary nitrate and sulfate PM that persist as a post-fog intense haze.
Contributions by these sources east of the Gorge contribute -57% of the haze on the eastern end of the
Gorge, with half of the nitrate and sulfate particulate from electric utility emissions and most of the rest
from transportation sources. Other sources contributing during the winter haze at the eastern end of the
December 2008
9-48
DRAFT-DO NOT QUOTE OR CITE
-------
1 Gorge are from sources outside the modeling domain (i.e. most of Washington and Oregon) and within
2 the Gorge (-23% and -10% respectively).
3 An assessment of concurrent measurements at the nearby Mt. Hood IMPROVE monitoring site
4 (45 km south of the Columbia River at 153 lm ASL), show that Columbia River Gorge haze levels and
5 especially the wintertime high nitrate/sulfate contributions to haze are not typical of the generally higher
6 elevation remote areas of the region (Pitchford et al., 2007). However the Gorge's high wintertime nitrate
7 and sulfate are found at the Hells Canyon IMPROVE site, which is similarly situated in a narrow canyon
8 of the Snake River almost 400 km east of the Gorge (from the VIEWS web site, see Table 9-1).
2000-2004 Baseline Average
20% Worst Days
IMPROVE Aerosol Extinction (Mm-1)
Ammonium Sulfate
Ammonium Nitrate
Organic Material
Elemental Carbon
Soil
Coarse Material
100
* <3 years
HI
Source: From the TSS web site.
Figure 9-31. Shows the IMPROVE monitoring sites in the WRAP region with at least three years of
valid data and identifies the six sites selected to demonstrate the apportionment
tools. Pie diagrams show the composition for the mean of the 20% worst haze
conditions.
December 2008
9-49
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Several example monitoring locations distributed across the northern and southern portions of the
western U.S. have been selected to illustrate the attribution results from the WRAP-sponsored attribution
analysis tools that estimate the relative responsibility for haze of the various PM species by source region
and source type. The selected sites include Olympic National Park (NP), WA; Yellowstone NP, WY; and
Badlands NP, SD across the north, and San Gorgonio Wilderness (W), CA; Grand Canyon NP, AZ and
Salt Creek W, NM across the south as shown in Figure 9-31.
WRAP-sponsored CMAQ modeling used virtual tracers of S02 and NOx emissions that tracked the
source region and category through the transport and transformation processes to particulate sulfate and
nitrate. This was used to produce pie diagrams of particulate sulfate and nitrate attribution results by
source region for each of these sites as shown in Figure 9-32 (produced using the TSS, see Table 9-1).
Based on these sites, over half of the particulate sulfate in remote areas of the Pacific coastal states is
from outside of the U.S. (Pacific offshore and outside of the domain). The outside of the domain values
were derived by simulating the fate of the boundary condition concentrations, which for the WRAP air
quality modeling were obtained using output from the GEOS-CHEM global air quality model (Fiore et
al., 2003). The sulfate fraction from the region labeled outside of domain was approximately uniform
throughout the western U.S. with site-to-site variation in the fraction caused mostly by variations in total
sulfate concentration. The more northerly sites have impacts from Canadian emissions, while the southern
sites have impacts from Mexican emissions. Half of the Salt Creek, New Mexico sulfate is from the
domestic source emissions further to the east, which also contribute about 20%-Badlands particulate
sulfate concentrations. A breakout of the emission sources from within the WRAP region by source type
(not shown) has most of the emissions from point sources, with the combination of motor vehicle, area
and wildfire emissions contributing from a few percent at the furthest eastern sites to about half at San
Gorgonio.
By comparison, the particulate nitrate is much more from domestic regional emission sources, with
~60%-~80% being from emissions within the WRAP region. For the west coast sites about 25% of the
nitrate is from a combination of Pacific offshore emissions (i.e. marine shipping) and outside domain
regions. Canadian emissions are responsible for about 10%-30% of the particulate nitrate for the three
northern sites, but Mexican emissions do not contribute appreciably to particulate nitrate for the three
southern sites. Motor vehicles are the largest contributing NOx source category responsible for particulate
nitrate for these six WRAP sites, with a combination of point, area and wildfire source categories also
contributing from about 10%-50% of the WRAP regional emissions.
December 2008
9-50
DRAFT-DO NOT QUOTE OR CITE
-------
2002
Sulfate 1 1 ua/m3
m
mvi
2002
Sulfate 0.6 ua/m3
\
2002
Sulfate 1.2 ua/m3
(a)
WRAP
Pacific Offshore
CENRAP
Eastern U.S.
Canada
Mexico
Outside Domain
2002
Sulfate 0.8 ua/m3
2002
Sulfate 0.5 ua/m3
2002
Sulfate 1.4 ua/m3
(b)
2002
Nitrate 1 6 ua/m3
2002
Nitrate 0.6 ua/nn3
2002
Nitrate 1.2 ua/m3
WRAP
Pacific Offshore
CENRAP
Eastern U.S.
Canai^a
~ Mexico
Outside Domain
2002
Nitrate 1.5 ua/nn3
2002
Nitrate 0.4 ua/m3
2002
Nitrate 0.5 ua/m3
Figure 9-32. Particulate sulfate (a) and nitrate (b) source attribution by region using CAMx
modeling for six western remote area monitoring sites: top left to right Olympic NP,
WA; Yellowstone NP, WY; Badlands NP, SD; bottom left to right San Gorgonio W, CA;
Grand Canyon NP, AZ; and Salt Creek W, NM. WRAP includes ND, SD, WY, CO, NM
and all states further west. CENRAP includes all states east of WRAP and west of the
Mississippi River including MN. eastern U.S. includes all states east of CENRAP. The
Pacific Offshore extends 300km to the west of CA, OR, and WA. Outside Domain refers
to the modeling domain, which extend hundreds of kilometer into the Pacific and
Atlantic Oceans and from Hudson Bay Canada to just north of Mexico City. This figure
was assembled from site-specific diagrams produced on the TSS web site (see Table
9-1).
WRAP only used the virtual tracer approach to investigate source locations and categories for SOi
and NOx emissions. A different type of virtual tracer modeling tool was used to track various the organic
carbon compounds and sort them into three groups. The first group labeled primary organics includes all
of the organics that are emitted directly as PM from any source type and location. The second group
December 2008
9-51
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
labeled anthropogenic secondary organics is PM produced in the atmosphere by aromatic VOCs. The
third category labeled biogenic secondary organics is PM produced in the atmosphere by biogenic VOCs.
Organic PM in the biogenic secondary category includes those that would functionally be considered
man-made emissions (e.g. those from agricultural crops and urban landscapes), though in most remote
areas of the west these man-made VOC emissions are small compared to those of the natural biogenic
sources. Figures 9-33, 9-34 and 9-35 show the monthly averaged apportionment of organic PM for the six
selected monitoring locations.
Organic Aerosol for All Days
Class I Area - Olympic NP, WA
3.20
^ 2.80
CO
f, =¦«
3
^ 2.00
o
¥ 1.60
ii,o
0
° 0.80
0.40
0.00
Organic Aerosol for All Days
Class I Areas - San Gorgonio W, CA: San Jacinto W, CA
2.80
©2.40
£
™2.00
1 160
I 1.20
3 0.80
0.40
0.00
| Anthro. Secondary Biogenic Secondary | Anthro. & Bio. Primary
Figure 9-33. Monthly averaged model predicted organic mass concentration apportioned into
primary and anthropogenic and biogenic secondary PM categories for the Olympic
National Park (top) and San Gorgonio Wilderness (bottom) monitoring sites. From the
TSS web site, see Table 9-1.
Based on the modeling results for these six sites and confirmed by measurements (see, e.g. Figure
9-10), a west to east decreasing gradient of organic mass exists with annual concentrations from ~2 (.ig/nr1
for the coastal state sites to ~1 |ig/nr for the intermountain west sites to less than 1 |ig/nr for the sites just
east of the Rocky Mountains, discounting the large fire impacts for July at Yellowstone N.P which raised
its annual mean to ~2 |ig/nr. At all of these remote-area sites anthropogenic secondary PM is estimated to
be a small fraction of the organic mass, with the largest fractional contribution at the San Gorgonio
55
55
Ml 1
n i i
i"i" i i
i
HI
(NCO^-mtOhwOOOTO*—
-------
1 monitoring site immediately downwind of the major Southern California urban areas, yet having less than
2 10% of the monthly mean organic mass from anthropogenic secondary PM. Of the six selected
3 monitoring sites, San Gorgonio has the highest fraction of the organic PM from primary emissions
4 (-57%), followed by Yellowstone (-55%), then the two eastern-most sites (Badlands -42% and Salt
5 Creek 41%), and with Grand Canyon and Olympic national parks the lowest fraction by primary
6 emissions (-37%). Yellowstone N.P would have had the lowest fraction of organic PM by primary
7 emissions had it not been for the month of July (the 11 month mean is 29%) when wild fire smoke
8 contributed.
14.00
12.00
: 10.00
jl
' 8.00
! 6.00
i 4.00
2.00
0.00
Organic Aerosol for All Days
Class I Areas - Grand Teton NP, VW: Red Rock Lakes NWRW, MT: Teton W, VVY: Yellowstone NP, WV
~~i i
! ! !_
kl") CD
CO CD O
Organic Aerosol for All Days
Hopi Point #1
| Anthro. Secondary
Biogenic Secondary
| Anthro. & Bio. Primary
Figure 9-34. Monthly averaged model predicted organic mass concentration apportioned into
primary and anthropogenic and biogenic secondary PM categories for the
Yellowstone National Park (top) and Grand Canyon (Hopi Point) (bottom) monitoring
sites. From the TSS web site, see Table 9-1.
December 2008
9-53
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
0.80
0.70
to 0.60
-I
3 0.50
| 0-40
| 0.30
c
3 0.20
0.10
0.00
Organic Aerosol for All Days
Class I Area - Salt Creek NWRW, NM
1.20
1.00
ci 0.80
3
C
| 0.60
"E
Hi
g 0.40
~
0.20
0.00
Figure 9-35. Monthly averaged model predicted organic mass concentration apportioned into
primary and anthropogenic and biogenic secondary PM categories for the Badland
National Park (top) and Salt Creek Wilderness (bottom) monitoring sites. From the
TSS web site, see Table 9-1.
Radiocarbon (14C) dating techniques were used to group ambient PM carbon into fossil and
contemporary source categories at 12 IMPROVE monitoring sites across the U.S., 8 of which are in the
WRAP region (Schichtel et al., 2008). Results of this work showed that contemporary carbon accounts for
about half the carbon in urban areas, 70%-97% in near-urban areas (i.e., San Gorgonio) and 82%-100% in
remote areas. Comparing these radiocarbon dating results with the WRAP virtual tracer modeling results
for organic aerosol (above), and presuming that the modeled anthropogenic secondary organic is fossil
carbon and the biogenic secondary is contemporary carbon, suggests that a large fraction of the model-
determined regional primary organic PM is from contemporary carbon sources (e.g. smoke from
wildfires).
Schichtel, et al. (2008) compared radiocarbon measurements at two sets of urban/rural paired sites
in the west (Mount Rainer/Seattle, and Tonto/Phoenix). Figure 9-36 shows that most of total carbon urban
excess (i.e. urban site concentration minus the regional site concentration) in the summer is from fossil
carbon sources (87% and 79% respectively), while in the winter there is a surprisingly high fraction of the
urban excess at both sites that is from contemporary carbon sources (41% and 47% respectively). This
implies that urban, and therefore anthropogenic, activities generate almost as much PM2 5 carbon from
Organic Aerosol for All Days
Class I Area - Badlands NP, SD
n_ i—i
¦|B|
i_i
CMfOT3-mCDh-COCnOT-CN
LnJ Li—in h J
HHW UUQ-«UJ n
T-(NCOT3-l/lCDh-COCnOT-(N
¦ Anthro. Secondary EH Biogenic Secondary Q Anthro. & Bio. Primary
December 2008
9-54
DRAFT-DO NOT QUOTE OR CITE
-------
contemporary sources (e.g. residential wood combustion) as from the fossil sources during the winter for
these two western urban areas.
6
5
n A
£ 4 -
en
5 3 H
"2
(5
O 1 -I
0
Summer
I
CO
10
II
CO
X
C
ro 2
. 03
O
(/>
CO
o
E
0
C
o
O
6
5
E 4
3 3
! 2 H
TO
° H
o
Summer
I
s2
° ®
o
ss
s
&
%
CO
E
g
cn
CO
en
ii
xs
¦
xs
Fossil
E
£
c
O
o
Winter
Winter
Source: (Schichtel et al., 2008)
Figure 9-36. Comparison of carbon concentrations between Seattle (Puget Sound site) and Mt.
Rainer (left) and between Phoenix and Tonto (right) showing the background site
concentration (gray) and the urban excess concentration (black) for total, fossil and
contemporary carbon during the summer and winter studies.
Contemporary carbon estimates for all of the IMPROVE network monitoring sites for data from
two summer seasons (June, July and August, 2004/2005) and two winter seasons (December, 2004/2005,
January and February, 2005/2006) were calculated from the measured EC/TC ratios using the 12-site
empirical relationship between radiocarbon determined contemporary carbon fraction and IMPROVE
measured EC/TC ratio (Schichtel et al., 2008). The results are displayed in color contour maps in Figure
9-37, which also shows the radiocarbon determined contemporary carbon for the 12 sites. The lowest
December 2008
9-55
DRAFT-DO NOT QUOTE OR CITE
-------
1 con tempo ran carbon estimates (< 60%) in both seasons are for urban areas. In the rural West, most of the
2 sites have over 90% of their PM carbon from contemporary carbon sources during the summer and from
3 60%-over 90% during the winter. In the rural East, most of the sites have 45%-90% of their PM carbon
4 from contemporary carbon sources during the summer and from 60%-over 90% in the winter.
Rocky
Mountains
Grand—T
Car yon /
nan r
Yoserrnle-T#
Great Smoky
Mountains
0.90
Tonto
0 84
[Fraction Contemporary|
¦ > 0.90 * 0.24
0.75 - 0,90 + 0.22
0.60 - 0.75 + 0.21
0.45 - 0.60 * 0.20
¦ <0.45 + 0.19
Mount
Ptwenix
0 56
Q Q
Briganline
0.79
O
Puget Sound
0 53
Proctor Mapla
0.87
Sugema
0.77
Tonto
0.71
Fraction Contemporary]
Briganline
0.78
O
Great Smoky
Mountains
0.82
O
Mount
Rainier
Phoenix
0.49
© O
B > 0.90 + 0.15
0.75 - 0.90 * 0.14
0.60 ¦ 0.75 * 0.13
0.45 - 0.60 + 0.12
|| < 0.45 + 0.11
Figure 9-37. Average contemporary fraction of PM2.5 carbon for the summer (top) and winter
(bottom), estimated from IMPROVE monitoring data (6/04 to 2/06) based on EC/TC
ratios. The contemporary values from radiocarbon dating for the 12 monitoring sites
are indicated in by colored circles with the site names. Color contours are shown to
aid in showing sites with similar values. Site locations are indicated by circles for
remote area sites and triangles for urban sites.
December 2008
9-56
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Schichtel, et al. (2008) showed a strong relationship between the site-averaged ratios of EC to total
carbon (EC/TC) and the site-averaged fraction of fossil carbon separately for the summer and winter data
sets (i.e. R2 of 0.71 and 0.87 respectively). Using regression analysis they estimated that the summer and
winter EC/TC ratios associated with purely fossil carbon were 0.35 ± 0.039 and 0.46 ± 0.028 respectively
and for purely contemporary carbon the EC/TC ratios were 0.12 ± 0.011 and 0.19 ± 0.0095. These ratios
are shown to be consistent with corresponding ratios from the literature for source testing primary fossil
and contemporary combustion sources respectively. They are also shown to be consistent with the 90
percentile value of the EC/TC ratio from the urban IMPROVE monitoring sites (0.41 and 0.44 for
summer and winter) and the 10th percentile values of the EC/TC ratio for remote areas IMPROVE
monitoring sites (0.07 and 0.16 for summer and winter), which they argue are dominated by fossil and
contemporary carbon respectively.
The largest sources of contemporary carbon are primary emissions from biomass burning and
secondary organic aerosol from biogenic precursor gases (e.g. terpenes from conifer forests). Schichtel, et
al. (2008) estimated the 12-site overall contribution by secondary organic PM to the summer
contemporary carbon fraction as 36 ± 6.4% by assuming the EC/TC ratio for contemporary carbon during
the winter represented the ratio of primary emissions only (i.e. no secondary organic PM formation in the
winter) and that the EC/TC ratio for primary emissions is independent of seasons. The same method
applied to the fossil carbon fraction yielded an estimate of 23 ± 10% of the fossil carbon PM from
secondary organic formation in the atmosphere during the summer. These estimates correspond to over
40% of the contemporary and over 35% of the fossil organic carbon being from secondary PM formation.
WRAP applied a weighted emissions potential analysis tool that combined gridded emissions data
with back-trajectory analysis that simulated the transport pathway to the various monitoring sites to infer
likely source region and emission categories for the 20% best and 20% worst haze conditions for each of
the IMPROVE PM speciation monitoring location in the West. Unlike the virtual tracer approach that
uses a full regional air quality simulation model, this method does not explicitly account for chemistry or
removal processes and it does not incorporate the sophisticated dispersion estimates (i.e., it uses one over
distance weighting for dispersion), so it should be considered a screening tool that has been found to be
helpful in identifying the likely sources contributing to haze. Primary organic and EC PM species results
from the weighted emissions potential tool for the worst 20% haze days using the 2000 to 2004 base years
emissions and trajectories, and the same trajectories with 2018-projected emissions for each of the six
selected western monitoring locations are shown in Figures 9-38 through 9-43.
December 2008
9-57
DRAFT-DO NOT QUOTE OR CITE
-------
100.00
90.00
80.00
70.00
60.00
o 50.00
-------
Sources and Areas of Potential Organic Carbon Emissions on Worst 20% Visibility Days
Class I Areas - San Gorgonio W, OA: San Jacinto W, OA
yu.uu
OII. II LI
7 U. U U
KIJ.UU
P 50.00
4 Li. IJ U
J U. U U
20.00
miJiJ
0.00
WRAP TSS - 3,01.¦are
¦
CA - 2000-04
PO - 2000-04
Sources and Areas of Potential Elemental Carbon Emissions on Worst 20% Visibility Days
Class I Areas - San Gorgonio W, CA: San Jacinto W, CA
90.00
ci U. U U
7II I III
60.00
2 50.00
40.00
-iU. UU
2U.UU
10.00
WRAP TSS - 3013TS
CA - 2000-04
PO - 2000-04
HWBDust I Road Dust | On-Road Mobile Q WRAP Area O&G ~ Biogenic HAnthroFire
Fugitive Dust Off-Road Mobile Q Off-Shore Q]Area Q Natural Fire | Point
Figure 9-39. Results of the weighted emissions potential tool applied to primary organic carbon
emissions (top) and EC emissions (bottom) for the baseline and projected 2018
emissions inventories for San Gorgonio W. Only source regions (WRAP states and
other regions) with the largest estimated contributions are shown (i.e., California and
Pacific Off-Shore from left to right). The scale is normalized (i.e., unitless) one over
distance weighted emissions multiplied by trajectory residence time. From the TSS
web site (see Table 9-1).
1 For Olympic N.P. (Figure 9-38), most of the primary organic as well as most of the EC PM is likely
2 to be from the state of Washington during the worst haze days. This is because the multi-day trajectories
December 2008
9-59
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
that transport emissions on its worst days tend to be short (within 200 km based on maps available on
TSS, see Table 9-1). Area sources, which include emissions from residential wood heating, watercraft,
non-mobile urban and other sources too small to be labeled as point sources, are the big contributors to
primary organic, while on- and off-road mobile emissions plus area sources are large contributors to the
EC at Olympic N.P. The 2018 projected growth in area sources and decrease in emissions of mobile
source emissions is anticipated to increase the haze by primary organic while reducing the haze by EC at
Olympic N.P. The same analysis applied to San Gorgonio (Figure 9-39), is similar in that the majority of
the emissions with the potential to contribute to primary organic and EC PM is from the home state,
California in this case. However, the likely importance of natural fire emissions for carbonaceous PM
species sites is substantially greater at San Gorgonio W. than it was for Olympic N.P.
The weighted emissions potential results applied to Yellowstone N.P. and Grand Canyon N.P.
(Figures 9-40 and 9-41) show the likely dominance of natural fire emissions in the intermountain western
U.S. to primary organic and EC PM. during worst haze conditions for these two locations. Numerous
states have emissions that have the potential to contribute noticeably to these carbonaceous species, due to
relatively long multi-day trajectories (500 km to 1000 km) on worst haze days, though for both sites the
home state has the greatest potential based on this inverse distance weighting approach. On- and off-road
mobile sources in Arizona and California have significant potential to contribute to Grand Canyon
carbonaceous particles, especially EC concentrations, probably due to some of the trajectories being over
the populated areas of these two states to the south and southwest of Grand Canyon.
For the most easterly of the selected WRAP sites, Badlands N.P. and Salt Creek, the weighted
emissions potential results for primary organic and EC (Figures 9-42 and 9-43) show potential
contributions from a greater number of states and multi-state regions than for selected sites further to the
west. This may be due in part to trajectories associated with worst haze conditions for these two sites
being moderately long (-500 km) and in multiple directions. Natural fire emissions have the greatest
potential to contribute to organic species PM at Badlands N.P, but are less likely to be dominant at Salt
Creek, or at either site in its contribution to EC PM concentrations. The contributions by emissions from
area and mobile sources from the home states and states to the east (Central States Regional Air
Partnership states are labeled CEN in the figures) are potentially greater than by natural fire; this is
especially true for contributions to EC PM.
December 2008
9-60
DRAFT-DO NOT QUOTE OR CITE
-------
Sources and Areas of Potential Organic Carbon Emissions on Worst 20% Visibility Days
Class I Areas - Grand Teton NP, WY: Red Rock Lakes NWRW, MT: Teton W, WY: Yellowstone NP, WY
70.00 —=
60.00
% 40.00
o
-------
Sources and Areas of Potential Organic Carbon Emissions on Worst 20% Visibility Days
Class I Area - Grand Canyon NP, AZ
50.00
1
££ 30.00 -H
UURAP TSS - 4/1OB0
Sources and Areas of Potential Elemental Carbon Emissions on Worst 20% Visibility Days
Class I Area - Grand Canyon NP, AZ
45.00 -H
40.00
35.00
^ 30.00 +
to 25.00
Q_
20.00 +
15.00
10.00 4-
5.00 -
0.00
UURAP TSS - 4/1/3E3
HWBDust | Road Dust | On-Road Mobile Q WRAP Area OSG D Biogenic | Anthro Fire
E Fugitive Dust D Off-Road Mobile [ Off-Shore QArea D Natural Fire | Point
Figure 9-41. Results of the weighted emissions potential tool applied to primary organic carbon
emissions (top) and EC emissions (bottom) for the baseline and projected 2018
emissions inventories for Grand Canyon N.P. Only source regions (WRAP states and
other regions) with the largest estimated contributions are shown (i.e., Arizona,
California, Mexico, New Mexico, Nevada, Oregon, Pacific Off-shore and Utah from left
to right). The scale is normalized (i.e., unitless) one over distance weighted emissions
multiplied by trajectory residence time. From the TSS web site (see Table 9-1).
December 2008
9-62
DRAFT-DO NOT QUOTE OR CITE
-------
30.00
27.00
24.00
21.00
18.00
-E
-------
Sources and Areas of Potential Organic Carbon Emissions on Worst 20% Visibility Days
Class I Area - Salt Creek NWRW, NM
40.00
36.00
32.00
28.00
24.00
20.00
16.00
12.00
8.00
4.00
0.00
= =
¦ i 1—i—
<
WRAP TSS - 3C31.3B3
36.00
33.00
30.00
24.00
^ 21.00
£? 18.00
ID
15.00
12.00
9.00
6.00
3.00
0.00
Sources and Areas of Potential Elemental Carbon Emissions on Worst 20% Visibility Days
Class I Area - Salt Creek NWRW, NM
r*
—
r==
In
~~
IAIRAP TSS - 3'31/3U5
¦WB Dust
Fugitive Dust
| Road Dust H On-Road Mobile | WRAP Area O&G Biogenic HAnthroFire
Off-Road Mobile [] Off-Shore Area ~ Natural Fire | Point
Figure 9-43. Results of the weighted emissions potential tool applied to primary organic carbon
emissions (top) and EC emissions (bottom) for the baseline and projected 2018
emissions inventories for Salt Creek W. Only source regions (WRAP states and other
regions) with the largest estimated contributions are shown (i.e., Arizona, California,
CenRAP, Colorado, eastern U.S., Idaho, Mexico, Montana, New Mexico, Nevada,
Oregon, Utah, and Wyoming from left to right). The scale is normalized (i.e., unitless)
one over distance weighted emissions multiplied by trajectory residence time. From
the TSS web site (see Table 9-1).
December 2008
9-64
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
WRAP applied the weighted emissions potential tool to assess likely source types and regions
contributing to coarse mass concentrations. The results for the six selected monitoring sites (not shown)
are as follows. Most dust at Olympic N.P. is likely to be from fugitive dust sources in Washington state,
while at San Gorgonio it is likely more from road dust with smaller amounts from fugitive dust sources.
The amount from wind-blown dust is small for both of these far westerly sites. Wind-blown dust is likely
the largest source contributing to coarse mass at Grand Canyon N.P., Badlands N.P. and Salt Creek
Wilderness with most of it originating in the home-state for those sites. The weighted emissions potential
results for coarse mass at Yellowstone are different from those of the other five selected sites in that Idaho
and Montana each have a higher potential to contribute to coarse mass on the worst haze days than the
home state (Wyoming), and that wind-blown and road dust both contribute to substantially as does
fugitive dust and natural fire.
In another WRAP-sponsored effort to better understand the causes of remote area haze in the
western U.S., each of the worst haze days for all western IMPROVE monitoring sites where dust (defined
as the sum of coarse mass and fine soil PM) was the largest contributor to light extinction was separately
assessed to categorize the most likely dust source (Kavouras et al., 2007) 2008 and the Causes of Haze
Website - see Table 9-1). Elemental composition was used to assess the likelihood that the dust was
associated with long-range transport from Asia. A regression analyses at each site between dust
concentrations and coincident local wind speed was used to generate site-specific estimates of local
windblown dust for each sample period. Finally, back trajectory analysis combined with maps constructed
of wind erosion potential (i.e., developed by combining soil types and land cover classifications) are used
in a manner similar to the weighted emissions potential analysis to identify the likelihood of regionally
transported wind-blown dust as the source. These assessments were conducted on each of the 610
so-called "worst dust haze days" at 70 monitoring sites for data from 2001 through 2003 to classify each
day by its likely contributions from Asian dust, local windblown dust, upwind transport and
undetermined. The undetermined category includes those sample periods that failed to be classified into
one of the other three source categories suggesting that mechanically suspended dust activities (e.g.
unpaved road dust, agricultural, construction and mining activities) may be responsible.
Of the 610 "worst dust haze days" at the 70 WRAP monitoring sites, 55 sample periods are
classified as Asian dust influenced, almost exclusively in the spring; 201 sample period are classified as
local windblown dust, mostly in the spring but some in all seasons; 240 sample periods are classified as
upwind transported dust, with a broader seasonal distribution centered on summer and few instances
during winter; and 114 are in the undetermined category with most seasonally distributed with most in
summer and least in winter. Most dust days occurred in the deserts of Arizona, New Mexico, Colorado,
western Texas and southern California, and these were dominated by local and regionally transported
December 2008
9-65
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
wind-blown dust (e.g. 84% for Salt Creek W.). Asian dust caused only a few of the worst dust days during
the 3-year assessment period, though it is an important source (i.e., 10% - 40% of the worst dust days) for
sites in the more northern regions of the West with greater vegetative land-cover where local and
regionally transported wind-blown dust was infrequent. The frequency of worst dust events classified as
undetermined was greatest for sites in the vicinity of large urban and agricultural areas such as those in
California and southern Arizona.
Source attribution of the particulate sulfate contribution to haze at Big Bend NP, TX was a primary
motivation for the BRAVO Study. Schichtel, et al. (2005) showed that during the four-month field
monitoring study (July through October, 1999) S02 emissions sources in the U.S. and Mexico were
responsible for -55% and -38% of the particulate sulfate respectively. Among U.S. source regions, TX
was responsible for -16%, eastern U.S. - 30%, and the western U.S. -9%. A large coal fired power plant,
the Carbon facility in Mexico, just south of Eagle Pass, TX, was responsible for about -19%, making it
the largest single contributor. Pitchford et al. (2005) put these results into the context of other component
contributions to regional haze, plus seasonal and longer-term variations in haze by particulate
components. Figure 9-44 shows the temporal variation of the contributions by the various S02 emissions
source regions plus the Carbon facility during the BRAVO Study period. The largest particulate sulfate
peak haze periods are dominated by infrequent large contribution by emission sources in TX and eastern
U.S., while Mexican sources including the Carbon facility are more frequent contributors to haze, but at
generally lower light extinction levels. Particulate nitrate contributions to haze at Big Bend NP are among
the lowest measured in the U.S. (-3% of light extinction on average and for worst haze episodes).
c
o
3
-Q
£
<
c
o
u
*
LU
O)
Sulfate Haze Source Attribution
n Carbon ¦ Other Mexico
EI Texas ¦ Eastern US
~ Western US C Other
Organics + LAC + Nitrates +
Fine Soil + Coarse
Clear Air
July 9
August 9
September 9
October 9
Source: Pitchford et al. (2005)
Figure 9-44. BRAVO Study haze contributions for Big Bend National Park, TX during a four-month
period in 1999. Shown are impacts by various particulate sulfate sources, as well as
the total light extinction level (black line) and Rayleigh or clear air light scattering.
December 2008
9-66
DRAFT-DO NOT QUOTE OR CITE
-------
' ACAD1
2004 Annual
ammN03f
Figure 9-45. Maps of spatial patterns for average annual particulate nitrate measurements (top),
and for ammonia emissions for April 2002 from the WRAP emissions inventory
(bottom).
December 2008
9-67
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
3.125
2.500
N02
1.875
1.500
1 5.000112
4.375
3.750
WRAP 36k base02a All Sources Emissions
2002 Yearly Total
1.875
P^E December 31,2002 0:00:00
mcnc Min= 0.000 at (4,1), Max= 4.995 at (130,68)
1 3.000112
2.625
2.250
WRAP 36k base02a All Sources Emissions
2002 Yearly Total
1.125
December 31,2002 0:00:00
Min= 0.000 at (4,1), Max= 4.041 at (130,68)
1.250
0.625
0.000 1
LOG(tons/yea|)
0.750
0.375
0.000 1
LOG(tons/yea|)
Figure 9-46. Maps of spatial patterns of annual NO (left) and NO2 (right) emissions for 2902 from
the WRAP emissions inventory.
Nitrate concentrations are a significant contributor to light extinction further to the north of Texas
in the center of the country. While sulfate can be in particulate form though not fully neutralized by
ammonia, nitric acid from NOx emissions requires neutralization by ammonium to become particulate
ammonium nitrate. One way to explore the causes of the Midwest nitrate bulge is to compare its spatial
distribution with the spatial distributions of NOx and ammonia emissions. Figure 9-45 shows a map of the
annual average particulate nitrate concentrations (top) with a map of ammonia emissions directly below
it. Animal agriculture is responsible for most of the ammonia emissions in the Midwest. The striking
similarity between the ambient particulate nitrate concentration and the ammonia emissions spatial
patterns with regional maximum centered on Iowa is in contrast to the NOx (i.e., NO + N02) emissions
spatial patterns, shown in Figure 9-46. NOx emissions are high over a broad region of the country
associated with the larger population densities and greater numbers of fossil fuel electric generation plant
generally to the east of the Midwest nitrate bulge. While both ammonia and nitric acid are needed to form
particulate ammonium nitrate, the maps suggest the Midwest nitrate bulge is due primarily to the
abundance of free ammonia (i.e. the amount beyond what is required to neutralize the acidic particulate
sulfate). By contrast the region to the east of the Midwest nitrate bulge should have plenty of nitric acid
given the higher emissions of NOx, but apparently has a deficiency of free ammonia. The few eastern
monitoring sites with locally high particulate nitrate (near southeastern PA) are located within a small
region of high density animal agricultural that shows up as a high ammonia emissions region in Figure 9-
45. Note that California's South Coast and Central Valley have both high ammonia and high NOx
emissions, explaining the high particulate nitrate contribution to haze there.
December 2008 9-68 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
To better understand the role of ammonia in the formation of the Midwest nitrate bulge, the
Midwest RPO and Central States Regional Air Partnership deployed a measurement program from late
2003 through early 2005 at 10 locations (9 rural and 1 urban) in the region (see Figure 9-47) to monitor
particulate sulfate, nitrate, and ammonium ions, plus the precursor gases sulfur dioxide, nitric acid, and
ammonia (Kenski, et al., 2004, Sweet, et al., 2005). These data have been used as input for
thermodynamic equilibrium modeling to assess the changes in PM levels that would result from changes
to precursor concentrations (Blanchard and Tanenbaum, 2005, Blanchard et al., 2007). Blanchard and
Tanenbaum (2005) and Blanchard et al. (2007) conclude that the current conditions at nine of the ten sites
are near the point of transition between the precursor species (nitric acid and ammonia) that limits the
formation of particulate nitrate. If excess ammonia increases, either by greater ammonia emissions or by
anticipated decreases in S02 emissions, then nitric acid levels would need to be reduced (via lower NOx
emissions) in order to reduce the particulate nitrate levels.
Given the comparability of particulate sulfate and nitrate with regard to their light extinction
efficiencies, their visibility impacts are proportional to the sum of their mass concentrations. A reduction
in sulfate caused by S02 emission reductions would reduce the particulate sulfate concentration, though
according to the thermodynamic equilibrium modeling for these sites the particulate nitrate concentration
will be increase somewhat. However the total particulate sulfate plus nitrate concentration would be
reduced so visibility impacts would be decreased. At current ammonium levels the predicted response of
changes to sulfate and nitric acid concentrations (i.e. S02 and NOx emissions changes) are similar in
respect to the resulting magnitude of changes to the total particulate sulfate plus nitrate concentrations. At
all but two sites the total particulate sulfate plus nitrate concentrations would decrease if either ammonia
or nitric acid where reduced.
Kenski et al., 2004
Figure 9-47. Midwest ammonia monitoring network.
December 2008
9-69
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
A further level of complications in understanding the response of particulate nitrate to changes in
precursor concentrations results from the temperature and humidity dependence of the partition between
particulate ammonium nitrate and the disassociated gaseous nitric acid and ammonia. This dependence
causes seasonal and even diurnal differences in the expected responses of particulate nitrate
concentrations to changes in precursor concentrations. As expected during the colder times of the year the
total particulate concentrations are more sensitive to changes in ammonia and nitric acid levels than
during the warmer seasons when sulfate levels are greater.
As shown in Figure 9-48, results of an air transport assessment to identify emission source areas
associated with high particulate nitrate at five monitoring locations in the East (four remote-area sites and
Toronto, Canada) implicate the high ammonia emissions region of the Midwest as a common source
region (Canada-U.S. Air Committee, 2004). This assessment does not preclude local sources of the
precursor gases responsible for particulate ammonium nitrate, but does suggest that long-range transport
of particulate nitrate or ammonia from the high emissions region of the Midwest is also contributing to
eastern nitrate episodes.
i> j 'J iUM >n i ii# F
Lye Brook VT.
_
Great Smoky Mountains NP
Upwind Probability Fields for Ammonium Nitrate
From Canada-U.S. Air Committee, 2004.
Figure 9-48. Upwind transport probability fields associated with high particulate nitrate
concentrations measured at Toronto, Canada; Boundary Water Canoe Area, MN;
Shenandoah National Park, VA; Lye Brook, VT; and Great Smoky Mountains National
Park, TN.
December 2008
9-70
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
In a similar airtransport assessment for measurements at Underhill, VT and at Rngantine. NJ,
Hopke et al. (2005) identified separate regions associated with particulate sulfate accompanied by trace
particulate components associated with coal burning (e.g. selenium) and accompanied by trace particulate
components associated with oil burning (e.g. vanadium). As shown in Figure 9-49, the coal-burning
related particulate sulfate for these two monitoring sites is associated with long-range transport from the
Ohio River Valley, while oil-burning related particulate sulfate is from more nearby emissions in the high
population region of coastal New York, New Jersey, Massachusetts, and Connecticut.
Hopke, etal. (2005).
Figure 9-49. Trajectory probability fields for periods with high particulate sulfate measured at
Underhill, VT and Brigantine, NJ (shown as white stars) associated with oil-burning
trace components (left) and with coal-burning trace components (right). Shown for
comparison are the interpolated SO2 emissions areal density contours for oil
combustion sources (emissions times 10) and coal combustion sources, displayed as
yellow and red contour lines respectively.
Hie Regional Aerosol Intensive Network (RAIN) was established by MANE-VU to generate
enhanced continuous visibility, plus fine particle mass and composition monitoring data at a string of
three monitoring locations along the transport path from the Ohio River Valley to coastal Maine
(NESCAUM, 2006). The dominant role of particulate sulfate in the northeast is well demonstrated by a
scatter plot of RAIN data that shows the relationship between particulate sulfate extinction, calculated
using the IMPROVE algorithm plotted against directly measured particle light scattering for hourly data
over a eight month period beginning in July 2004 at the Acadia National Park, ME monitoring site (see
Figure 9-50). Particulate sulfate explains 90% of the variance in particulate light scattering even though it
December 2008
9-71
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
is responsible for only about 64% of the total light extinction (annual averaged value from the VIEWS
web site). Adding the contribution by the second-largest regional contributor to light extinction,
particulate organic carbon with about 14%, does not improve the variance explained, but does increase
the slope to 0.78. The noticeable difference between these two plots is that the particulate sulfate alone
underestimates light scattering during low haze periods (points on the plot are below the regression line
for light scattering < 70Mm_1), while the agreement is improved with the addition of particulate organic
carbon contributions to haze (regression slope is nearer to one and reduced bias for low haze periods).
Particulate nitrate contribution to light extinction at Acadia is about 10% on average.
Nephelometer Bsp versus SULFATE Bs
Nephelometer Bsp vs (SULFATE + OMC) Bs
y = 0.64X - 1.87
Ft2 = 0.90
n = 4098
y = 0.78x + 4.57
R2 = 0.89 Oy ^
n = 1627 / ^
©
o-'o
o
0
I SO 100 150 200 250 300 3;
2-Hour Nephelometer Bsp (1/Mmeters)
2-Hr Nephelometer (1/Mmeters)
Source: RAIN Preliminary Data Analysis Report, NESCAUM, 2006.
Figure 9-50. Scatter plots of particulate sulfate (left) and particulate nitrate and organic mass
(right) versus nephelometer measured particle light scattering for Acadia National
Park, ME.
Particulate nitrate levels are considerably lower in the sulfate-dominated warmer southeastern U.S.
than in the Northeast and upper Midwest. Blanchard, et al. (2007) conducted thermodynamic equilibrium
modeling on data from the eight Southeastern Aerosol Research and Characterization (SEARCH)
monitoring sites and found that total particulate nitrate plus sulfate is much more responsive to changes in
sulfate concentrations than to changes in nitric acid concentrations, which in turn is more responsive than
changes in ammonia concentrations.
December 2008
9-72
DRAFT-DO NOT QUOTE OR CITE
-------
Great Smoky Mtns, TN (20% Worst Days)
¦K
a>
CD
<
-12.
¦10.
-8.
-6.
-4,
-2.
0.
00 T
00
00
00
00
00
00
O 4./- o
^*
V
JJ'
CD
< "1
0.
1.
00
00
00
00
00
.00
.00
.00
*ov
&
r&'
<£>' <£>'
~ Bio VOC
~ Anthro VOC
¦ BCs
~ MRPO
~ M-VU
~ CEN
¦ VISTAS
~ WV
~ VA
~ TN
~ SC
~ NC
~ MS
~ KY
~ GA
~ FL
~ AL
Source: NCDENR, 2007.
Figure 9-51. CMAQ air quality modeling projections of visibility responses on the 20% worst haze
days at Great Smoke National Park, NC (top) and Swanquarter Wilderness, NC
(bottom) to 30% reductions from a projected 2009 emission inventory of
visibility-reducing pollutants by source category and geographic areas.
1 The VISTAS RPO commissioned an emissions sensitivity study using CMAQ modeling on winter
2 and summer 2009 emissions projected from the 2002 emissions inventory (NCDENR, 2007). Figure 9-51
December 2008
9-73
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
contains bar plots for two North Carolina class I areas that indicate projected changes in light extinction
for the worst haze day due to 30% emissions reductions by particulate species, source types and location
across the Southeastern states modeling domain (i.e. as far west as Texas, as far north as Pennsylvania, as
far south as the Florida Keys and as far east as -300 km from the North Carolina coast). Great Smoky
Mountains, in the southern Appalachian Mountains has the greatest sensitivity to changes by S02
emissions from electrical generation units (EGU) and to a lesser extent other S02 emission sources in the
region. Reductions of NOx emissions from ground or point sources are not nearly as effective as S02
reductions in reducing the light extinction levels at Great Smoky Mountains. This is due principally to the
worst days at Great Smoky Mountains occurring during the summer, when temperatures are too high to
support high particulate nitrate concentrations. For the same reason, ammonia emission reductions are
also ineffective. Swanquarter Wilderness, NC is a coastal location where some of the worst haze days are
during the winter and include contributions from particulate ammonium nitrate. Both S02 and ammonia
emissions reductions would be effective at reducing worst haze days at the Swanquarter Wilderness,
though NOx emissions are not as effective presumably because the atmosphere is ammonia-limited for
particulate nitrate production.
9.3.5. Urban Visibility Valuation and Preference
The Clean Air Act §302(h) defines public welfare to include the effects of air pollution on
"... visibility, ... and personal comfort and wellbeing. " Though good visibility conditions in Class I (e.g.
National Parks) and wilderness areas have long been recognized as important to the public welfare (see
discussions in EPA (2004 and 2005) and Chestnut and Dennis (1997), visibility conditions in urban areas
also contribute to the public welfare. Visibility impairment may be caused by either natural or manmade
conditions (or both), but it is only impairment that occurs as a result of air pollution (either alone or in
combination with water vapor or other atmospheric conditions) that can be mitigated by regulations such
as the Regional Haze Rule (40 CFR 51.300 through 309) or the Secondary National Ambient Air Quality
Standards (NAAQS). Visibility impairment resulting from air pollution is referred to as visual air quality
(VAQ).
Visibly poor air quality causes people to be concerned about substantive health risks, but degraded
VAQ adversely affects people in additional ways. These include the aesthetic benefits of better visibility,
improved road and air safety, and enhanced recreation in activities like hiking and bicycling. Because the
human health impacts of air pollution are regulated under the Primary NAAQS, it is necessary to separate
out the aesthetic and wellbeing components associated with the visibility condition produced by a given
level of air pollution when assessing the need for additional regulation to protect the public welfare effect
December 2008
9-74
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
of visibility under the Secondary NAAQS. The degree to which previous human preference and valuation
studies for VAQ have adequately made this distinction and separation is an important issue in applying
results from available studies in a Secondary NAAQS (or benefits estimation for any policy effecting
VAQ) context. The remainder of this discussion is focused on those aesthetic and wellbeing qualities
associated with a given VAQ in urban areas.
The term "urban visibility" is used to refer to VAQ throughout a city or metropolitan area. Urban
visibility includes the VAQ conditions in all locations that people experience in their daily lives, including
scenes such as residential streets and neighborhood parks, commercial and industrial areas, highway and
commuting corridors, central downtown areas, and views from elevated locations providing a broad
overlook of the metropolitan area. Thus urban visibility, which is sometimes referred to as 'residential
visibility,' encompasses more than the visibility conditions only at an individual's specific place of
residence, but all the VAQ they see on a regular basis. Urban visibility includes not only major cities, but
VAQ conditions in smaller towns and cities. The key distinction is between visibility conditions in urban
and suburban locations and visibility in rural or wilderness settings such as the Class 1 areas defined by
the Clean Air Act, which include National Parks and similar natural settings.
Visibility has direct significance to people's enjoyment of daily activities and their overall
wellbeing. Visibility conditions can be described both as an aesthetic quality as well as a scientifically
measurable set of atmospheric conditions. Due to the subjective nature of aesthetics, people's preferences
with respect to visibility are difficult to express or quantify, but people have expressed in many different
ways that they enjoy and value a clear view. A number of social science disciplines have undertaken to
link perceived urban visibility to an array of effects reflecting the overall desire for good VAQ, and the
benefits of improving currently degraded VAQ. This wide range of diverse studies have identified types of
benefits of good VAQ in addition to those directly connected with air-pollution related health effects such
as respiratory diseases and premature mortality.
For example, psychological research has demonstrated that people are emotionally affected by
VAQ such that their overall sense of wellbeing is diminished (e.g., Bickerstaff and Walker, 2001).
Researchers have also shown that perception of pollution is correlated with stress, annoyance, and
symptoms of depression (Bickerstaff and Walker, 2001; Evans and Jacobs, 1982; Jacobs et al., 1984).
Sociological research has demonstrated that VAQ is deeply intertwined with a "sense of place," effecting
people's sense of the desirability of a neighborhood quite apart from the actual physical conditions of the
area (e.g., ABT, 2002; Day, 2007; Elliott et al., 1999; Howel et al., 2002). Public policy research finds that
people think it is important to protect visibility, and accept the concept of setting standards to protect
visibility (e.g., ABT, 2001; BBC Research & Consulting, 2002; Ely et al., 1993; Pryor, 1996). Finally,
economic valuation research has measured the amount of money that people are willing to pay to protect
December 2008
9-75
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
or improve both urban visibility (e.g., summary review in Beron et al., 2001; Chestnut and Dennis, 1997)
and natural locations such as National Parks and other locations defined by the Clean Air Act as Class I
visibility areas (e.g., summary review in Chestnut and Dennis, 1997).
The purpose of the remainder of this section is to review four urban preference studies, as well as
one new urban visibility valuation study not previously discussed in an EPA Criteria Document or
OAQPS Staff Paper. This literature is relevant to the review of a Secondary NAAQS standard concerning
VAQ, as well as a review of potentially including urban visibility valuation in a damage function
approach (separately estimating individual effect categories) an economic benefit analysis.
Urban visibility has been examined in two types of studies directly relevant to the NAAQS review
process: urban visibility preference studies and urban visibility valuation studies. Both types of studies are
designed to evaluate individuals' desire (or demand) for good VAQ where they live, using different
metrics to evaluate demand. Urban visibility preference studies examine individuals' demand by
investigating the basic question "what level of visibility degradation is unacceptable," while economic
studies examine demand by investigating "how much would you be willing to pay to improve visibility."
9.3.5.1. Urban Visibility Preference Studies
One group of urban visibility research projects focused on identifying preferences for urban VAQ
without necessarily estimating the economic value of improving visibility. This group of preference
studies used a common focus group method to estimate the level of visibility impairment that respondents
described as "acceptable." The specific definition of acceptable was largely left to each individual
respondent, allowing each to identify their own preferences.
There are three completed studies that used this method, and one additional pilot study (designed as
a survey instrument development project) that provided additional information (Table 9-2). The
completed studies were conducted in Denver, Colorado (Ely et al., 1991), two cities in British Columbia,
Canada (Pryor, 1996), and Phoenix, Arizona (BBC Research & Consulting, 2002). The pilot study was
conducted in Washington, DC (ABT, 2001).
Each study collected information in a focus group setting, presenting slides depicting various
visibility conditions. All four studies used photographs of a single scene from the study's city; each photo
included images of the broad downtown area and spreading out to the hills or mountains composing the
scene's backdrop. The maximum sight distance under good conditions varied by city, ranging from 8
kilometers in Washington, DC to mountains hundreds of kilometers away in Denver. Multiple photos of
the same scene were used to present approximately 20 different levels of visibility impairment. The
Denver and British Columbia studies used actual photographs taken in the same location to depict various
visibility conditions. The Phoenix and Washington, DC pilot study used photographs prepared using the
December 2008
9-76
DRAFT-DO NOT QUOTE OR CITE
-------
1 WinHaze software from Air Resource Specialists (ARS). WinHaze is a computer-imaging software
2 program that simulates visual air quality differences of various scenes, allowing the user to "degrade" an
3 original near-pristine visibility condition photograph to create a photograph of each desired VAQ level.
Table 9-2. Summary of urban visibility preference studies.
Denver, CO
Phoenix, AZ
2 British Columbia cities
Washington, DC (pilot)
Report Date
1991
2003
1996
2001
Duration of session
45 min
50 mins
2 h
Compensation
None (civic groups)
$50
None (class room exercise)
$50
# focus group
17
27 total at 6 locations,
4
1
sessions
Including 3 in Spanish
# participants
214
385
180
9
Age range
adults
18-65+
University students
27-58
Annual or seasonal
Wintertime
Annual
Summertime
Annual
# total scenes
presented
Single scene of
downtown with
mountains in background
Single scene of downtown
and mountains, 42 km
maximum distance
Single scene from each city
Single scene of DC Mall and
downtown, 8 km maximum
sight
# of total visibility
conditions presented
20 levels (+ 5 duplicates)
21 levels (+ 4 duplicates)
20 levels (10 each from each city)
20 levels (+ 5 duplicates)
Source of slides
Actual photos taken
between 9am and 3pm
WinHaze
Actual photos taken at 1pm or 4pm
WinHaze
Medium of
presentation
Slide projection
Slide projection
Slide projection
Slide projection
Ranking scale used
7 point scale
7 point scale
7 point scale
7 point scale
Visibility range
presented
11 to 40 dV
15 to 35 dV
13 to 25 dV (Chilliwack) 13.5 to 31.5
dV (Abbotsford)
9 to 38 dV
Health issue
directions
Ignore potential health
impacts; visibility only
Judge solely on visibility, do
not consider health
Judge solely on visibility, do not
consider health
Health never mentioned,
"Focus only on visibility"
Key Questions asked
a) Rank VAQ (1-7 scale)
a) Rank VAQ (1-7 scale)
a) Rank VAQ (1-7 scale)
a) Rank VAQ (1-7 scale)
b) Is each slide
"acceptable"
b) Is each slide "acceptable"
b) Is each slide "acceptable"
b) Is each slide "acceptable"
c) "How much haze is
too much?"
c) How many days a year
would this picture be
"acceptable"
c) if this hazy, how many
hours would it be acceptable
(3 slides only)
d) valuation question
Mean dV found
"acceptable"
20.3 dV
23 to 25 dV
-23 dV(Chilliwack),
-19 dV(Abbotsford)
-20 dV (range 20-25)
4 A common characteristic of the three visibility preference studies was that each were conducted in
5 the West where distant mountains were shown in the photograph used to elicit local participant responses
6 about visibility. Among other issues, the Washington D.C. pilot study was the first step in a process to
7 expand the results to other regions where typical scenes may have different sensitivity to perceived
December 2008
9-77
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
visibility changes in PM air quality and where participants may have different acceptable visibility
preference levels.
One notable finding of the three visibility preference studies and the one pilot study is the general
degree of consistency in the median preferences for an acceptable level of visibility degradation. The
range of median acceptable preference level from the four studies is 19 to 25 deciviews (dV), the
preferred measure of visibility impairment. Measured in terms of visual range (VR), these median
acceptable levels are between 30 and 55 km.
9.3.5.2. Denver, Colorado Urban Visibility Preference Study
The Denver urban visibility preference study (Ely et al., 1993) was conducted on behalf of the
Colorado Department of Public Health and Environment (CDPHE). The study conducted a series of focus
group sessions with 17 civic and community groups in which a total of 214 individuals were asked to rate
slides. The slides depicted varying levels of VAQ for a well-known Denver vista, including a broad view
of downtown Denver with the mountains to the west composing the scene's background. The participants
were instructed to base their judgments on three factors:
1. the standard was for an urban area, not a pristine national park area where the standards might
be more strict;
2. the level of an urban visibility standard violation should be set at a VAQ level considered to
be unreasonable, objectionable, and unacceptable visually; and
3. judgments of standards violations should be based on visibility only, not on health effects.
Participants were shown 25 randomly ordered slides of actual photographs. The visibility
conditions presented in the slides ranged from 11 to 40 dV, approximating the 10th to 90th percentile of
wintertime visibility conditions in Denver. The participants rated the 25 slides based on a scale of 1 (poor)
to 7 (excellent), with 5 duplicates included. They were then asked to judge whether the slide would
violate what they would consider to be an appropriate urban visibility standard (i.e., whether the level of
impairment was "acceptable" or "unacceptable"). The individual's judgment of a slide's VAQ and
whether the slide violated a visibility standard were highly correlated (Pearson correlation coefficient
greater than 80%), as were the VAQ ratings and the yes/no "acceptable" response. The participant's
median response was that a visibility level of 20.3 dV (extinction coefficient bext = 0.76/km, or VR ~ 51
km) was judged as "acceptable." The CDPHE subsequently established a Denver visibility standard at
this level (defined as bext = 0.76/km), based on the median 50% acceptability findings from the study.
December 2008
9-78
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
9.3.5.3. Phoenix, Arizona Urban Visibility Preference Study
The Phoenix urban visibility preference study (BBC Research & Consulting, 2002) was conducted
on behalf of the Arizona Department of Environmental Quality. The Phoenix study patterned its focus
group survey process after the Denver study. The study included 385 participants in 27 separate focus
group sessions. Participants were recruited using random digit dialing to obtain a sample group designed
to be demographically representative of the larger Phoenix population. Focus group sessions were held at
six neighborhood locations throughout the metropolitan area to improve the participation rate. Three
sessions were held in Spanish in one region of the city with a large Hispanic population (25%), although
the final overall participation of native Spanish speakers (18%) in the study was modestly below the
targeted level. Participants received $50 as an inducement to participate.
Participants were shown a series of 25 images of the same vista of downtown Phoenix, with South
Mountain in the background at a distance of about 40 km. Photographic slides of the images were
developed using WinHaze. The visibility impairment levels ranged from 15 to 35 dV (the extinction
coefficient, bext, range was approximately 45/km to 3.5/km, or a visual range of 87 to 12 km). Participants
first individually rated the randomly shown slides on a VAQ scale of 1 (unacceptable) to 7 (excellent).
Participants were instructed to rate the photographs solely on visibility, and to not base their decisions on
either health concerns or what it would cost to have better visibility. Next, the participants individually
rated the randomly ordered slides as "acceptable" or "unacceptable," defined as whether the visibility in
the slide is unreasonable or objectionable. Better visibility conditions (15 dV and 20 dV) were judged
"acceptable" by 90 percent of all participants. At 24 dV nearly half of all participants thought the VAQ
was "unacceptable," with almost three-quarters judging 26 dV as unacceptable.
The Phoenix urban visibility study formed the basis of the decision of the Phoenix Visibility Index
Oversight Committee for a visibility index for the Phoenix Metropolitan Area (Arizona Department of
Environmental Quality, 2003). The Phoenix Visibility Index establishes an indexed system with 5
categories of visibility conditions, ranging from "Excellent" (14 dV or less) to "Very Poor" (29 dV or
greater). The "Good" range is 15 to 20 dV. The environmental goal of the Phoenix urban visibility
program is to achieve continued progress through 2018 by moving the number of days in lower quality
categories into better quality categories.
9.3.5.4. British Columbia, Canada Urban Visibility Preference Study
The British Columbia urban visibility preference study (Pryor, 1996) was conducted on behalf of
the Ministry of Environment. The study conducted focus group sessions that were also developed
following the methods used in the Denver study. Participants were students at the University of British
December 2008
9-79
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Columbia, who participated in one of four focus group sessions with between 7 and 95 participants. A
total of 180 respondents completed surveys (29 did not complete the survey).
Participants in the study were shown slides of two suburban locations in British Columbia:
Chilliwack and Abbotsford. Using the same general protocol as the Denver study, Pryor found that
responses from this study found the acceptable level of visibility was 23 dV in Chilliwack and 19 dV in
Abbotsford. Pryor (1996) discusses some possible reasons for the variation in standard visibility
judgments between the two locations. Factors discussed include the relative complexity of the scenes,
potential bias of the sample population (only University students participated), and the different levels of
development at each location. Abbotsford (population 130,000) is an ethnically diverse suburb adjacent to
the Vancouver Metro area, while Chilliwack (population 70,000) is an agricultural community 100 km
east Vancouver in the Frazier Valley.
The British Columbia urban visibility preference study is being considered by the B.C. Ministry of
the Environment as a part of establishing urban and wilderness visibility goals in British Columbia.
9.3.5.5. Washington, DC Urban Visibility Pilot Preference Study
The Washington, DC urban visibility pilot study (Abt Associates 2001) was conducted on behalf of
the EPA, and was designed to be a pilot focus group study, an initial developmental trial run of a larger
study. The intent of the pilot study was to study both focus group method design and potential survey
questions. Due to funding limitations, only a single focus group session was held, consisting of one
extended session with 9 participants. No further urban visibility focus group sessions were held in
Washington, DC.
Due to the small number of participants, it is not possible to make statistical inferences about the
opinions of the general population. The study does, however, provide additional useful information about
urban visibility studies, potentially helping to both better understand previous studies as well as design
future studies.
The study also adopted the general Denver study method, modifying it as appropriate to be
applicable in an eastern urban setting which has substantially different visibility conditions than any of
the three western locations of the other preference studies. Washington's (and the entire East) visibility is
typically substantially worse than western cities, and has different characteristics. Washington's visibility
impairment is primarily a uniform whitish haze dominated by sulfates, relative humidity levels are higher,
the low lying terrain provides substantially shorter maximum sight distances, and many residents are not
well informed that anthropogenic emissions impairs visibility on hazy days.
The Washington focus group session included questions on valuation, as well as on preferences.
The focus group was asked to state their preferences measured in an increase in the general cost of living
December 2008
9-80
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
for certain levels of improvement in visibility on a typical summer day. A general cost of living approach
is one payment vehicle approach that can be used in willingness to pay studies, especially for
environmental issues arising from multiple diverse emission sources (e.g., transportation, electricity
generation, industry, etc.) making a specific price increase potentially misleading.
The first part of the focus group session was designed to be an hour long, and was comparable to
the focus group sessions in the Denver and Phoenix studies. A single scene was used; a panoramic shot of
the Potomac River, Washington mall and downtown Washington, DC. In the first part of the session
people were asked to rate the VAQ of 25 photographs (prepared using WinHaze, and projected on a large
screen), judge the acceptability of visibility level in each slide, and answer the valuation questions. The
second half of the session, however, was a moderated discussion session about the format and content of
the first phase of the session. In this moderated discussion, participants were asked about their
understanding of each question asked in the first half of the session. Particular issues in designing a focus
group session were also explored. Important participant comments included:
1. Participants had been asked how they reacted to the initial direction to base their answers
only on visibility, but health was never explicitly mentioned by the focus group moderator.
Participants strongly agreed with the decision to not mention that health effects are associated
with visibility impairment. They understood the directions as meaning they should ignore
health issues, and said their answers would have been different if they included health as well
as visibility in their judgments.
2. Differentiating between haze and weather conditions was difficult. Weather was not discussed
in the focus group session, and the photographs were WinHaze altered photos with identical
weather conditions. Participants mentioned they were still confused about the role of weather
and humidity in the different visibility conditions presented in the photos.
3. Questions about how many hours an impairment level would be acceptable were confusing.
Most participants were normally indoors during most of the day, so questions about duration
of outdoor conditions were difficult to answer.
4. Participants strongly agreed that not mentioning the purpose of the study, or the sponsor, until
the very end (after all the questions were answered) was viewed as very important. Most felt
this information would have influenced their answers.
December 2008
9-81
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
9.3.5.6. Urban Visibility Valuation Studies
The economic importance of urban visibility has been examined by a number of studies designed to
quantify the benefits (or willingness to pay) associated with potential improvements in urban visibility.
Urban visibility valuation research prior to 1997 was summarized in Chestnut and Dennis (1997), and
was also described in the 2004 Air Quality for PM (p. 4-186 to 4-190, (U.S. EPA, 2004) and the 2005
OAQPS PM NAAQS Staff Paper (EPA, 2005). These reviews summarize 34 estimates (based on different
cities or model specifications) from six different studies. Since the mid 1990s, however, only one new
valuation study of urban visibility has been published.
One urban visibility benefit assessment not included in those reviews is "The Benefits of Visibility
Improvement: New Evidence from the Los Angeles Metropolitan Area" (Beron et al., 2001). Rather than
a contingent valuation method (CVM) technique used in the majority of other urban visibility valuation
studies, Beron et al. used a housing market hedonic technique. The housing hedonic methods were used in
previous urban visibility studies by Murdoch and Thayer, 1988, and Trijonis et al., 1985. A housing
market hedonic study views a housing unit as composed of a bundle of attributes, and uses housing sale
price data from a large number of units in a metropolitan area to estimate the value of each component.
Hedonic pricing has been used to estimate economic values for environmental effects that have a direct
effect on housing market values. It relies on the measurement of differentials in property values under
various environmental quality conditions including air pollution, visibility and other environmental
amenities such as access to nearby beaches and parks, as well as by physical attributes of the house and
attributes of the neighborhood.
Beron et al. (2001) obtained data on approximately 840,000 owner-occupied, single family housing
sales between 1980 and 1995 from the California South Coast Air Basin (composed of Los Angeles and
Orange Counties, and the portions of Riverside and San Bernardino Counties in the greater metropolitan
area). The real estate data included information on the sale price of the house, 13 housing attributes
(square footage, number of bathrooms, etc.), 9 neighborhood attributes (percent poverty, school quality,
FBI crime index, etc.), and three air pollution variables: ozone, particulates (measured by total suspended
particulates, or TSP), and visibility. Visibility was measured as the annual average of visual range,
measured in miles, and was obtained from seven airports within the study region. The visibility range was
from 12.4 miles (Los Angeles International Airport, 1991) to 31.9 miles (Palm Springs Airport, 1995).
Ozone data (39 monitors) and TSP data (40 monitors) were obtained from the South Coast Air Quality
Management District. Annual mean values for each year were calculated for ozone and TSP.
Beron et al. (2001) presented results for a hypothetical basin-wide 20% visibility improvement, or
an increase from 15.3 to 18.4 miles, which is equivalent to approximately 27.6 deciViews (dV, a preferred
measure of visibility) to 25.8 dV. The initial results reflect the change in the purchase price of a house
December 2008
9-82
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
associated with this difference in VAQ, which can be interpreted as a present value of a stream of annual
values over the lifetime of the house. The authors therefore selected a time horizon (30 years) and an
interest rate (8%) to calculate an annual per household benefit per dV ranging from $484 to $1,756. The
Beron results are higher than the CVM-based values summarized in Chestnut and Dennis (1997), which
ranged from $12 to $132 per dV. It should be noted that the $132 CVM values cited by Chestnut and
Dennis (1997) is from a study in the Los Angeles area (Brookshire, 1979). The Beron et al. (2001) results
are also higher than the Trijonis et al. (1990) hedonic study in the Los Angeles area, which had a range of
$134 to $360 per dV. All values reported here are in terms of 1994 prices.
A critical question for all urban visibility valuation studies is the extent to which the estimated
values strictly reflect preferences for visibility, and do not include a component of preferences for
reducing health risk from air pollution. The ability to isolate the value of visibility from within the
collection of intertwined benefits from visual air quality, which is inherently multi-attributed, is a
challenge for all visibility valuation studies. Each study attempts to isolate visibility from other effect
categories, but different studies take different approaches.
Beron et al. (2001) include two measures of air pollution directly related to health effects in their
housing market hedonic study, ozone and particulates (using TSP as the metric for particulates), as well as
visibility. They argue that the presence of the two health-related pollution levels results in a estimated
hedonic demand function for visibility that successfully separates the health component of demand for
overall air quality from the visibility component. An alternative interpretation is that the estimated
visibility function still includes a component of health risk because the housing market data does not
support completely isolating the demand for visibility (due to correlated variables, omitted variables,
measurement error, model specification error, etc.) from demand for health risk reductions measured by
the two health related air quality metrics.
A key issue in interpreting the Beron et al. (2001) results is whether the objective measures of air
quality characteristics (e.g., visibility, PM concentrations, etc.) capture people's perceptions of the
different aspects of air quality in a given location. To the extent the people simultaneously use what they
see regarding VAQ as an indicator of the overall air quality including potential health risks, then including
all the measures in the equation is not necessarily sufficient to isolate one effect from the other.
9.4. Deposition of PM
Airborne particles, their gas-pahse precursors, and their transformation products are removed from
the atmosphere by wet and dry deposition processes. These deposition processes transfer PM pollutants to
December 2008
9-83
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
other environmental media where they can alter the structure, function, diversity, and sustainability of
complex ecosystems.
9.4.1. Forms of Deposition
9.4.1.1. Fine vs. Coarse PM
Research summarized by the previous NAAQS PM assessment illustrated the complexity of
deposition processes in patchy forested landscapes and the effects of vertical stratification within
canopies. There are also differences in the deposition behavior of fine and coarse particles. Coarse
particles generally settle nearer their site of formation than do fine particles. In addition, the chemical
constitution of individual particles is correlated with size. For example, much of the base cation and
heavy metal burden is present on coarse particles.
Fine PM may act as a carrier for materials such as herbicides that are phytotoxic. Fine PM provides
much of the surface area of particles suspended in the atmosphere, whereas coarse PM provides much of
the mass of airborne particles. Surface area can influence ecological effects associated with the oxidizing
capacity of fine particles, their interactions with other pollutants, and their adsorption of phytoactive
organic compounds. Fine and coarse particles also respond to changes in atmospheric humidity,
precipitation, and wind, and these can alter their deposition characteristics.
Fine PM is often a secondary pollutant that forms within the atmosphere, rather than being directly
emitted from a pollution source. It derives from atmospheric gas-to-particle conversion reactions
involving nucleation, condensation, and coagulation, and from evaporation of water from contaminated
fog and cloud droplets. Fine PM may also contain condensates of volatile organic compounds, volatilized
metals, and products of incomplete combustion, including polycyclic aromatic hydrocarbons (PAH) and
BC (soot) (U.S. EPA, 2004).
Coarse PM is mainly a primary pollutant, having been emitted from pollution sources as fully
formed particles derived from abrasion and crushing processes, soil disturbances, desiccation of marine
aerosol emitted from bursting bubbles, hygroscopic fine PM expanding with humidity to coarse mode,
and/or gas condensation directly onto preexisting coarse particles. Suspended primary coarse PM may
contain iron, silica, aluminum, and base cations from soil, plant and insect fragments, pollen, fungal
spores, bacteria, and viruses, as well as fly ash, brake lining particulates, debris, and automobile tire
fragments. Coarse mode particles can be altered by chemical reactions and/or physical interactions with
gaseous or liquid contaminants.
December 2008
9-84
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Direct and indirect radiative effects of atmospheric particulates influence local and regional
climate. In particular, dust derived from wind erosion can be lifted to high altitude and be transported long
distances from the source location (Mahowald et al., 2002). This process is most pronounced in desert
regions. Desert dust is the main atmospheric aerosol component in many arid and semi-arid regions,
especially the Sahara and the desert regions of central Asia (Zakey et al., 2006b). In recent years, a
number of efforts have been made to simulate the desert dust cycle in global climate models (c.f., Cakmur
et al., 2004; Cakmur et al., 2006; Miller et al., 2004; Zakey et al., 2006a; Zender et al., 2004).
Exposure to a given mass concentration of PM may lead to widely differing phytotoxic and other
environmental outcomes depending upon the particular mix of PM constituents involved. Especially
important in this regard are S and N components of PM, which are addressed in the Integrated Science
Assessment for Oxides of Nitrogen and Sulfur (NOxSOx ISA), and effects of particulate heavy metals
and organic contaminants. This variability has not been characterized adequately. Though effects of
specific chemical fractions of PM have been described to some extent, there has been relatively little
research aimed at defining the effects of unspeciated PM on plants or ecosystems.
9.4.1.2. Deposition Modes
Wet Deposition
Wet deposition results from the incorporation of atmospheric particles and gases into cloud droplets
and their subsequent precipitation as rain or snow, or from the scavenging of particles and gases by
raindrops or snowflakes as they fall (Lovett, 1994). Wet deposition depends on precipitation amount and
ambient pollutant concentrations. Receptor (i.e., vegetation) surface properties have little effect on wet
deposition, although leaves can retain liquid and solubilized PM. In terrain containing extensive
vegetative canopies, any material deposited via precipitation to the upper stratum of foliage is likely to be
intercepted by several foliar surfaces before reaching the soil. This allows such processes as foliar uptake,
chemical transformation, and re-suspension into the atmosphere to occur.
Landscape characteristics can affect wet deposition via orographic effects and by the closer
aerodynamic coupling to the atmosphere of tall forest canopies as compared to the shorter shrub and
herbaceous canopies. Following wet deposition, humidity and temperature conditions further affect the
extent of drying versus concentrating of solutions on foliar surfaces, which influence the rate of metabolic
uptake of surface solutes (Swietlik and Faust, 1984). The net consequence of these factors on direct
physical effects of wet deposited PM on leaves is not known (U.S. EPA, 2004).
Rainfall introduces new wet deposition and also redistributes throughout the canopy previously
dry-deposited PM (Peters and Eiden, 1992). Both effects contribute to the relationships between canopy
December 2008
9-85
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
leaf area and foliar contact and influence the potential direct PM effects on vegetation. The concentrations
of suspended and dissolved materials are typically highest at the onset of precipitation and decline with
duration of individual precipitation events (Hansen et al., 1994). Sustained rainfall removes much of the
accumulation of dry-deposited PM from foliar surfaces, reducing direct foliar effects and combining the
associated chemical burden with the wet deposited material (Lovett, 1994) for transfer to the soil. Intense
rainfall may contribute substantial total particulate inputs to the soil, but it also removes bioavailable or
injurious pollutants from foliar surfaces. This washing effect, combined with differential foliar uptake and
foliar leaching of different chemical constituents of PM, alters the composition of the rainwater that
reaches the soil and the pollutant burden that is taken up by plants. Once in the soil, chemical particle
constituents may affect biogeochemical cycles of major, minor, and trace elements. Low intensity
precipitation events, in contrast, may deposit significantly more particulate pollutants to foliar-surfaces
than high intensity precipitation events. Additionally, low-intensity events may enhance foliar uptake
through the hydrating of some previously dry-deposited particles (U.S. EPA, 2004).
Dry Deposition
Dry particulate deposition, especially of heavy metals, base cations, and organic contaminants, is a
complex, poorly characterized process. It appears to be controlled primarily by such variables as
atmospheric stability, macro- and micro-surface roughness, particle diameter, and surface characteristics
(Hosker and Lindberg, 1982). The range of particle sizes, the diversity of canopy surfaces, and the variety
of chemical constituents in airborne PM have made it difficult to predict and to estimate dry particulate
deposition (U.S. EPA, 2004).
Dry deposition of atmospheric particles to plant and soil surfaces affects all exposed surfaces.
Larger particles >5 (j,m diameter are dry deposited mainly by gravitational sedimentation and inertial
impaction. Smaller particles, especially those with diameters between 0.2 and 2 |im. are not readily
dry-deposited and may travel long distances in the atmosphere until their eventual deposition, most often
via precipitation. Plant parts of all types, along with exposed soil and water surfaces, receive steady
deposits of dry dusts, EC, and heterogeneous secondary particles formed from gaseous precursors
(U.S. EPA, 1982).
Estimates of regional particulate dry deposition infer fluxes from the product of variable and
uncertain particulate concentrations in the atmosphere and even more variable and uncertain measured or
modeled estimates of dry Vd parameterized for a variety of specific surfaces (e.g., Brook et al., 1999).
Even for specific sites and well defined particles, uncertainties are large. Modeling the dry deposition of
particles to vegetation is at a relatively early stage of development, and it is not currently possible to
identify a best or most generally applicable modeling approach (U.S. EPA, 2004).
December 2008
9-86
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Occw/f Deposition
The occurrence of occult deposition tends to be more restricted geographically, mainly to coastal
and high mountain areas. Several factors make occult deposition particularly effective, where it occurs,
for the delivery of dissolved and suspended particulates to vegetation. Concentrations of
particulate-derived materials are often many-fold higher in cloud or fog water than in precipitation or
ambient air due to orographic effects and gas-liquid partitioning. In addition, fog and cloud water deliver
chemical particle constituents in a bioavailable hydrated form to foliar surfaces. This enhances deposition
by sedimentation and impaction of submicron aerosol particles that exhibit low Vd prior to fog droplet
formation (Fowler et al., 1989). Deposition to vegetation in fog droplets is proportional to wind speed,
droplet size, concentration, and fog density. In some areas, typically along foggy coastlines or at high
elevations, occult deposition represents a substantial fraction of total deposition to foliar surfaces (Fowler
et al., 1991).
9.4.2. Methods for Estimating Dry Deposition
Methods for estimating dry deposition of particles are more restricted than for gaseous species and
fall into two major categories: surface analysis methods, which include all types of measurements that
examine contaminant accumulation on surfaces of interest; and atmospheric deposition rate methods,
which measure contaminants in the atmosphere and on surfaces from which one may estimate the
deposition rate (Davidson and Wu, 1990). Surface extraction or washing methods characterize the
accumulation of particles on natural receptor surfaces of interest or on experimental surrogate surfaces.
These techniques rely on methods designed specifically to remove only surface-deposited material. Total
surface rinsate may be equated to accumulated deposition or to the difference in concentrations in rinsate
between exposed and control (sheltered) surfaces and may be used to refine estimates of deposition.
Foliar extraction techniques may underestimate deposition to leaves because of uptake and translocation
processes that remove pollutants from the leaf surface (Garten and Hanson, 1990; Taylor Jr. et al., 1988).
Foliar extraction methods also cannot distinguish gas- from particle-phase sources (Bytnerowicz et al.,
1987a; Bytnerowicz et al., 1987b; Dasch, 1987; Lindberg and Lovett, 1985; Van Aalst, 1982).
The National Dry Deposition Network (NDDN) was established in 1986 to document the
magnitude, spatial variability, and trends in dry deposition across the U.S. Currently, the network operates
as a component of the Clean Air Status and Trends Network (CASTNet) (Clarke et al., 1997). A
significant limitation on current capacity to estimate regional effects of PM is inadequate knowledge of
the mechanisms and factors governing particle dry deposition to diverse surfaces (U.S. EPA, 2004).
December 2008
9-87
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Dry deposition can not be directly measured. Deposition rates and totals are often calculated as the
product of measured ambient concentration and a modeled deposition velocity. This method is widely
used because atmospheric concentrations are easier to measure than are dry deposition rates, and models
have been developed to estimate deposition velocities. Ambient pollutant concentrations and
meteorological conditions required for application of inferential models are routinely collected at
CASTNet dry deposition sites. Monitored chemical species are limited to ozone, sulfate, nitrate,
ammonium, sulfur dioxide, and nitric acid. The temporal resolution for the ambient concentration
measurements and dry deposition flux calculations is hourly for ozone and weekly for the other chemical
substances (Clarke et al., 1997).
Collection and analysis of stem flow and throughfall can also provide useful estimates of
particulate deposition when compared to directly sampled precipitation. The method is most precise for
PM deposition when gaseous deposition is a small component of the total dry deposition and when
leaching or uptake of compounds of interest out of or into the foliage is not a significant fraction of the
depositional flux total because these lead to positive and negative affects in the calculated totals.
Foliar washing, whether using precipitation or experimental lavage, is one of the best available
methods to determine dry deposition of PM to vegetated ecosystems. Major limitations include the site
specificity of the measurements and the restriction to elements that are largely conserved within the
vegetative system. Surrogate surfaces have not been found that can adequately replicate essential features
of natural surfaces; and therefore do not produce reliable estimates of particle deposition to the landscape.
Micrometeorological methods employ eddy covariance, eddy accumulation, or flux gradient
protocols for quantifying dry deposition. These techniques require measurements of PM concentrations
and of atmospheric transport processes. They are currently well developed for ideal conditions of flat,
homogeneous, and extensive landscapes and for chemical species for which accurate and rapid sensors
are available. Additional studies are needed to extend these techniques to more complex terrain and more
chemical species.
The eddy covariance technique measures vertical fluxes of gases and fine particles from
calculations of the mean covariance between the vertical component of wind velocity and pollutant
concentration (Wesely et al., 1982). This technique is limited by its requirement for sensors capable of
acquiring concentration data at 5 to 20 Hz. For the flux gradient or profile techniques, vertical fluxes are
calculated from a concentration difference and an eddy exchange coefficient determined at discrete
heights (Erisman et al., 1988; Huebert et al., 1988). Most measurements of eddy transport of PM have
used chemical sensors (rather than mass or particle counting) to focus on specific PM components. These
techniques have not been well developed for generalized particles and may be less suitable for coarse
particles that are transported efficiently in high frequency eddies (Gallagher et al., 1988).
December 2008
9-88
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Emissions of PM and deposition are generally highest in urban areas as a result of industrial
processes, vehicular traffic and home heating (Lu et al., 2003; Rocher et al., 2004). Urban settings
therefore continue to be a major focus of atmospheric particulate research. Previous work tended to focus
on the size of particles as a consequence of epidemiologic studies that found strong associations between
particle size and respiratory illness. More recent studies have also assessed the chemical composition of
particles and processes involved in their formation. Gilli et al. (2008) found that concentrations of
secondary particles, those formed in the atmosphere from gaseous phase pollutants or resulting from
adsorption of elements to emitted or resuspended particles, could comprise up to 45% of PMi0 in Italian
cities, but that this percentage could vary considerably depending on local conditions such as surrounding
landscape, climate, meteorological conditions, and characteristics of urban development. The most
important secondary components were N03 and S042 . although CI" was also important in coastal
settings. Secondary components tend to be more important constituents of the small fractions of PM (less
than 10 |im) than the larger fractions. This characteristic is relevant to abatement strategies because the
small fraction of PM will often increase the most during high pollution episodes (van Dingenen et al.,
2004). In the cities studied by Gilli et al. (2007) the PM25 fraction comprised about 60% of the total PMi0.
Table 9-3. Factors potentially important in estimating mercury exposure and how they are
addressed in this study.
Factor Importance and Possible Effect on Mercury Exposure
Type of anthropogenic source of mercury Different combustion and industrial process sources are anticipated to have different local scale impacts due to physical
source characteristics (e.g., stack height), the method of waste generation (e.g., incineration or mass burn) or mercury
control devices and their effectiveness.
Mercury emission rates from stack Increased emissions will result in a greater chance of adverse impacts on environment.
Mercury species emitted from stack More soluble species will tend to deposit closer to the source.
Form of mercury emitted from stack Transport properties can be highly dependent on form.
Deposition differences between vapor Vapor-phase forms may deposit significantly faster than particulate-bound forms,
and particulate-bound mercury
Transformations of mercury after Relatively nontoxic forms emitted from source may be transformed into more toxic compounds,
emission from source
Transformation of mercury in watershed Reduction and revolatilization of mercury in soil limits the buildup of concentration,
soil
Transport of mercury from watershed Mercury in watershed soils can be a significant source to water bodies and subsequently to fish,
soils to water body
Transformation of mercury in water body Reduction, methylation, and demethylation of mercury in water bodies affect the overall concentration and the MHg
fraction, which is bioaccumulated in fish.
Facility locations Effects of meteorology and terrain may be significant.
Location relative to local mercury source Receptors located downwind are more likely to have higher exposures. Influence of distance depends on source type.
Contribution from non-local sources of Important to keep predicted impacts of local sources in perspective,
mercury
Uncertainty Reduces confidence in ability to estimate exposure accurately.
Source: Modified from U.S. EPA (1997)
December 2008
9-89
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
The most useful records of long-term atmospheric metal deposition are recorded as accumulations
in ice, snow, peat, and lake sediment. Case studies presented by Norton (2007) focused on three elements
(Cd, Pb, Hg), which are biologically active, have negative consequences for ecosystem and human health,
and are dominated by atmospheric inputs. Sedimentary pollution records suggest that atmospheric
deposition of these elements in the U.S. peaked in about the period 1965 to 1975, but subsequently
declined by 75% or more. High concentrations still reside in soil in some areas, but the flux through
aquatic ecosystems has decreased in recent years (Norton, 2007).
Few studies in the past have reconstructed, from lake sediment records, the atmospheric
depositional history of trace metals and PAHs in lakes adjacent to coal-fired power plants. However,
Donahue et al. (2006) analyzed sediment from Wababun Lake, which is located in Alberta, Canada in
proximity (within 35 km) to four power plants built since 1950. Trace metal concentrations of Hg, Cu, Pb,
As, and Se in lake sediment increased by 1.2- to 4-fold. The total PAH flux to surface sediments was 730
to 1100 |ig/m2/yr. which was two to five times higher than in two lakes situated 20 km to the north and 70
km to the south.
The U.S. EPA (1997) compiled an assessment of the sources and environmental effects of Hg in the
U.S. A variety of factors were found to influence Hg deposition, fate and transport (Table 9-3).
9.4.3. Factors Affecting Dry Deposition
In the size range ~0.1 to 1.0 |im. where Vd is relatively independent of particle diameter (Figure 9-
52), particulate deposition is controlled by roughness of the surface and by the stability and turbulence of
the atmospheric surface layer. Impaction and interception dominate over diffusion as dry deposition
processes, and the Vd is considerably lower than for particles that are either smaller or larger than this size
range (Shinn, 1978).
Deposition of particles between 1 and 10 |im diameter is strongly dependent on particle size
(Shinn, 1978). Larger particles within this size range are collected more efficiently at typical wind speeds
than are smaller particles (Clough, 1975), suggesting the importance of impaction. Impaction is related to
wind speed, the square of particle diameter, and the inverse of receptor diameter as a depositing particle
fails to follow the streamlines of the air in which it is suspended around the receptor. When particle
trajectory favors a collision, increasing either wind speed or the ratio of particle size to receptor cross
section increases the probability of collision.
December 2008
9-90
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
0)
E
u
u
O
5
£
O
(A
O
D-
0)
O
1,000 -
100 —
10 —
1 —
0.1 —
0.01 —
0.001 —
0.000
•¦¦¦•¦¦(inn 5lokes Law
¦ ¦¦¦¦¦ Brownian Diffusion
• Pelersand Eiden (1992)
i Little and Wiffen (1977)
I I I
0.001 0.01 0.1 1
Particle Diameter (pm)
T~
10
100
Source: U.S. EPA (2004).
Figure 9-52. The relationship between particle diameter and deposition velocity for particles.
Values measured in wind tunnels by Little and Wiffen (1977) over short grass with
wind speed of 2.5 m/s closely approximate the theoretical distribution determined by
Peters and Eiden (1992) for a tall spruce forest. These distributions reflect the
interaction of Brownian diffusivity (descending dashed line), which decreases with
particle size and sedimentation velocity (ascending dotted line from Stokes Law),
which increases with particle size. Intermediate-sized particles (0.1 to 1.0 |jm) are
influenced strongly by both particle size and sedimentation velocity, and deposition is
independent of size.
Empirical measurements of Vd for fine particles under wind tunnel and field conditions have often
been several-fold greater than predicted by available theory (Unsworth and Wilshaw, 1989). A large
number of transport phenomena, including streamlining of foliar obstacles, turbulence structure near
surfaces, and various phoretic transport mechanisms are not well characterized (U.S. EPA, 2004). The
discrepancy between measured and predicted values of Vd may reflect model limitations or experimental
limitations in the specification of the effective size and number of receptor obstacles. Available reviews
(e.g., U.S. EPA, 1996, 2004) suggest the following generalizations: (1) particles >10 |im exhibit variable
Vd between 0.5 and 1.1 cm/s depending on friction velocities, whereas a minimum particle Vd of 0.03
cm/s exists for particles in the size range 0.1 to 1.0 pm; (2) the Vd of particles is approximately a linear
December 2008
9-91
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
function of friction velocity; and (3) deposition of particles from the atmosphere to a forest canopy is
from 2 to 16 times greater than deposition in adjacent open terrain like grasslands or other low vegetation.
9.4.3.1. Leaf Surface Effects on Deposition Velocity
The chemical composition of PM is not usually considered to be a primary determinant of Vd.
Rather, the plant leaf surface has an important influence on the Vd of particles, and therefore on the rate
and total of dry deposition to the terrestrial environment. Relevant leaf surface properties include
stickiness, microscale roughness, and cross-sectional area. These properties affect the probability of
impaction and particle bounce. The efficiency of deposition to vegetation also varies with leaf shape.
Particles impact more frequently on the adaxial (upper surface) surface than on the abaxial (lower
surface). Most particles accumulate in the midvein, central portion of leaves. The greatest particle loading
on dicotyledonous leaves is frequently on the adaxial surface at the base of the blade, just above the
petiole junction. Precipitation washing probably plays an important role in this distribution pattern
(U.S. EPA, 2004).
Lead particles have been shown to accumulate to a greater extent on older than younger needles
and twigs of white pine, suggesting that wind and rain may be insufficient to fully wash the foliage.
Fungal mycelia (derived from windborne spores) were frequently observed in intimate contact with other
particles on leaves, which may reflect minimal re-entrainment of the spore due to shelter by the particles,
mycelia development near sources of soluble nutrients provided by the particles, or simply co-deposition
(Smith and Staskawicz, 1977).
Leaves with complex shapes tend to collect more particles than do those with more regular shapes.
Conifer needles are more efficient than broad leaves in collecting particles by impaction, reflecting the
small cross section of the needles relative to the larger leaf laminae of broadleaves and the greater
penetration of wind into conifer canopies than broadleaf ones (U.S. EPA, 2004).
9.4.3.2. Canopy Surface Effects on Deposition Velocity
Surface roughness increases particulate deposition, and Vd is usually greater for a forest than for an
nonforested area and greater for a field than for a water surface. Different size particles have different
transport properties and Vd. The upwind leading edges of forests, hedge rows, and individual plants are
primary sites of coarse particle deposition. Impaction at high wind speed and the sedimentation that
follows the reduction in wind speed and carrying capacity of the air in these areas lead to preferential
deposition of larger particles (U.S. EPA, 2004).
December 2008
9-92
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Air movement is slowed in proximity to vegetated surfaces. Canopies of uneven age or with a
diversity of species are typically aerodynamically rougher and receive larger inputs of dry-deposited
pollutants than do smooth, low, or monoculture vegetation (Garner et al., 1989; U.S. EPA, 2004).
Canopies on slopes facing the prevailing winds receive larger inputs of pollutants than more sheltered,
interior canopy regions.
All foliar surfaces within a forest canopy are not equally exposed to particle deposition. Upper
canopy foliage tends to receive maximum exposure to coarse and fine particles, but foliage within the
canopy tends to receive primarily fine aerosol exposures. The dry deposition of fine-mode particles and
unreactive gases tends to be more evenly distributed throughout the canopy.
Both uptake and release of PM constituents can occur within the canopy. The leaf surface is a
region of leaching and uptake. Exchange also occurs with epiphytic organisms and bark and through
solubilization of previously dry deposited PM. Vegetation emits a variety of particles and particulate
precursor materials.
9.4.4. Magnitude of Dry Deposition
Dry deposition of PM is most effective for coarse particles. These include primary geologic
materials and elements such as iron and manganese. By contrast, wet deposition is more effective for fine
particles of secondary atmospheric origin and elements such as cadmium, chromium, lead, nickel, and
vanadium (Reisinger, 1990; Smith, 1990; U.S. EPA, 2004). The relative magnitudes of the different
deposition modes varies with ecosystem type, location, elevation, and chemical burden of the atmosphere
(U.S. EPA, 2004).
High-elevation forests generally receive larger particulate deposition loadings than equivalent low
elevation sites. Higher wind speeds at high elevation enhance the rate of aerosol impaction. Orographic
effects enhance rainfall intensity and composition and increase the duration of occult deposition.
High-elevation forests are often dominated by coniferous species with needle-shaped leaves that enhance
impaction and retention of PM delivered by all three deposition modes.
Urban stormwater can be rich in heavy metals and other contaminants derived from atmospheric
deposition, and can be a major source of pollutant inputs to water bodies in urban settings. Urban
stormwater runoff can also be toxic to aquatic biota, partly due to trace metal concentrations (Greenstein
et al., 2004; Sabin et al., 2005; Schiff et al., 2002). These processes are largely a function of the
impervious nature of much of the ground surface in urban areas (i.e., buildings, roads, sidewalks, parking
lots, construction sites). Dry-deposited pollutants can build up, especially in arid and semi-arid
environments, and then be washed into surface waters with the first precipitation event.
December 2008
9-93
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
9.4.4.1. Using Vegetation for Estimating Atmospheric Deposition
Whereas direct real-time measurement of deposition or air concentrations of atmospheric
contaminants is desirable, it is not always practical (Howe et al., 2004). Instead, passive time-integrative
methods are frequently used. These can involve analysis of vegetative tissues as a record of pollutant
exposure, or analysis of lake sediment cores or ice cores to determine changes in pollutant input over
time. There is a general assumption that the concentration of an analyte in vegetation reflects the
time-integrated concentration of that analyte in the air. The development of deposition layers in sediment
or ice cores allows the possibility of determining the effects of changes in the atmospheric concentration
over periods of years, decades, or longer.
Biomonitoring methods are important in air pollution assessment and provide a complement for
more typical instrumental analyses. It is well know that mosses can accumulate heavy metals to high
levels in response to atmospheric deposition. The effects of deposited metals on the mosses have been less
well studied. Tremper et al. (2004) exposed mosses of two species to roadside conditions and sampled
them over a period of three months. Under field conditions, chlorophyll concentrations in moss tissue
were not affected by metal contamination and accumulation.
Because mosses accumulate dissolved and PM deposited from the atmosphere, they have been used
extensively in Europe as surrogate collectors for estimating bulk (wet plus dry) deposition of metals. The
ease and low cost of this method has enabled regional assessments to be conducted throughout Europe.
Despite its wide use, however, several papers have pointed out complications in the use of mosses
to quantify metal deposition rates. Zechmeister (1998) found that the uptake efficiency for 12 heavy
metals in 3 species of moss was similar, but that uptake efficiency in a fourth species was uncorrelated
with the other species for about half the metals considered. Zechmeister (1998) also showed that
productivity of an individual species can vary greatly among sites. To calculate atmospheric deposition of
metals from accumulation in mosses, both the metal concentration and the rate of biomass production is
needed. Further complication was shown in the study of Shakya et al. (2008), which revealed that
accumulation of Cu, Zn and Pb decreased chlorophyll content. Sites with higher deposition levels may
therefore have lower rates of productivity than cleaner sites.
Differences in uptake efficiencies among species and productivity among sites has led to the use of
a single moss species placed in mesh bags that can be distributed to areas where that species of moss does
not grow naturally. Studies to standardize this passive deposition monitoring approach have been limited.
Adamo et al. (2007) evaluated the effects of washing with water, oven drying, and acid washing as
preteatments and found little difference in uptake efficiencies, although the ratio of the collecting surface
area to mass was found to be a key factor in uptake efficiency.
December 2008
9-94
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Couto et al. (2004) investigated dry versus bulk deposition of metals using transplanted moss bags.
This study showed that at some sites dry deposition exceeded bulk deposition, a likely outcome of
wash-off of dry deposited particles. This study also documented intercationic displacement and leaching
as a result of acidic precipitation. The authors concluded that the accumulated metal concentration
represented an unstable equilibrium between inputs and outputs of elements that were a function of the
local environment and weather during the exposure period. They also concluded that it was not possible
to extrapolate calibrations between metal accumulation in moss and atmospheric deposition of metals to
areas with different weather conditions, precipitation pH, and air contaminant concentrations.
Zechmeister et al. (2003) also presented results demonstrating the problems with dry deposited particles
that can be washed off by rain. These studies indicate that moss is not a completely effective collector of
total particle deposition. Deposition estimates from moss accumulation probably represent values that fall
between wet deposition and total deposition.
A European moss biomonitoring network has been in place since 1990 (Harmens et al., 2007).
Sampling surveys are repeated every five years. The survey conducted in 2005/2006 occurred in 32
countries at over 7000 sites. The network reports metal concentrations associated with live moss tissue.
Trends analysis of these data showed statistically significant decreases over time in moss concentrations
for As, Cu, V, and Zn. Trends were not observed for Cr, Fe, or Ni. Results for individual countries
participating in the survey have also been published. In Hungary, major pollution sources were readily
detected by moss sampling (Otvos et al., 2003). Somewhat higher metal concentrations in mosses in 1997
than in other European countries were attributed to the use of a different moss species in the Hungarian
survey (Otvos et al., 2003). Similar sampling in Romania showed regions with contamination that were
among the highest in Europe. These results were consistent with known air quality problems in Romania
(Lucaciu et al., 2004). Because particulate deposition is not well characterized using this method, spatial
patterns and temporal trends for particulate metal deposition in Europe only provide crude estimates of
relative deposition patterns.
The use of moss to assess heavy metal deposition has received much less attention in the U.S. than
in Europe. A study conducted in the Blue Ridge Mountains, VA, found that metal concentrations in moss
were related to elevation and canopy species at some sites (Schilling and Lehman, 2002). However, metal
concentrations in moss were not related to concentrations in the O horizon of the soil. Other measurement
methods for trace metal deposition were not available to compare with moss concentrations.
Epiphytic lichens have also been used to evaluate heavy metal accumulation. Helena et al. (2004)
found substantially increased concentrations of metals in lichens transplanted from a relatively clean
region to an area in proximity to a metal smelter. The presence of specific species of bryophyte or lichen
can serve as an effective bioindicator of metal contamination (Cuny et al., 2004). In some studies, tree
December 2008
9-95
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
bark has been used as a biomonitor for atmospheric deposition of heavy metals (Baptista et al., 2008;
Pacheco and Freitas, 2004; Rusu et al., 2006).
Biomonitoring using mosses, lichens, or other types of vegetation has been well established as a
means of identiff spatial patterns in the atmospheric deposition of heavy metals in relation to power
plants, industry, and other point and regional emissions sources. More recently, a number of studies
(Lopez Alonso et al., 2002; 2003a; 2003b) have used cattle reared predominantly on local forage as a
means of monitoring atmospheric inputs of Cu, Ar, Zn, and Hg. For example, Hg emissions from coal
fired power plants in Spain had a substantial effect on Hg accumulation by calves (Lopez Alonso et al.,
2003a). Accumulation of Hg by cattle extended to about 140 to 200 km downwind from the source.
Yang et al. (2007a) investigated the effectiveness of pine needles as passive air samplers for
semi-volatile organic compounds (SOCs), such as PAHs, that are partially or completely
particle-associated in the atmosphere. PAH distribution patterns are complicated by their properties,
which span a broad range of octanol-air partition coefficients. This allows them to be present in both
vapor and particle phases. In addition, the air-plant partitioning of PAHs is affected by air temperature and
atmospheric stability (2007a). DeNicola et al. (2005) documented the suitability of a Mediterranean
evergreen oak (Quercus ilex) to serve as a passive biomonitor for atmospheric contamination with PAH in
Italy.
9.4.4.2. Deposition to Canopies
Tree canopies have been shown to increase dry deposition from the atmosphere, including
deposition of PM. Dry deposition rates in the canopy are commonly estimated by the difference between
throughfall deposition and deposition measured by an open collector, although the use of this approach to
specifically quantify particulate deposition is complicated by gaseous deposition to leaf surfaces and, for
some elements, leaching and uptake. Avila and Rodrigo (2004) found that trace metal deposition in
throughfall in a Spanish oak forest were higher than bulk deposition for Cu, Pb, Mn, V, and Ni, but not for
Cd and Zn. This study also found that dry deposition of Cu, Pb, Zn, Cd and V occurred, but that canopy
uptake of Zn and Cd also occurred. Leaching of Mn and Ni from the foliage was observed as well.
Leaching of Ni, Cu, Mn, Rb, and Sr from a red spruce-balsam fir canopy by acidic cloud water was also
measured in a study by Lawson et al. (2003). These studies suggest that leaching of trace metals from
forest canopies varies with tree species and the acidity of precipitation. Throughfall therefore cannot be
assumed to represent total deposition of heavy metals without evaluating uptake and leaching at the
specific study site.
Physical models have provided an alternative to estimating dry deposition to canopies with
throughfall measurements. Recently, Pryor (1996) and Binkowski (2004) identified an additional
December 2008
9-96
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
complication in that models typically hold particle size constant. Nevertheless, there may be significant
modification of particle size distributions during the deposition process. Condensation processes in the
vicinity of the canopy can increase particle size and may explain discrepancies between observations and
modeled dry deposition that is based on air sampling of particulates above the canopy.
The use of pine and oak canopies as bioindicators of atmospheric trace metal pollution was
investigated by Aboal et al. (2004). Metal concentrations in leaves were found to be one to three orders of
magnitude lower than in mosses collected in this study. As an ecosystem pool, metals in leaves were
likely to be much more important than those in mosses. The authors concluded, however, that these tree
species were not effective bioindicators of atmospheric deposition of heavy metals.
The effectiveness of tree canopies in capturing particulates was investigated as a method for
improving air quality by Freer-Smith et al. (2004). This study showed that with consideration of planting
design, location of pollution source, and tree species, planting of trees can be affective at reducing
particulate air pollution. However, this approach does not address the possible effects of the captured
pollution on trees, soils and surface waters.
An important aspect of global Hg cycling is the extent to which the Hg stored in forest vegetation
originates from the soil or the atmosphere. In other words, do plants recycle Hg by uptake from the soil
and then return it to the soil in litterfall, or do plants directly capture atmospheric Hg and then deliver it to
the soil as an external source? The question was addressed in a mesocosm experiment by Ericksen et al.
(2003). Aspen trees were grown in gas-exchange chambers in Hg-enriched soil (12.3 ±1.3 jxg/g) and the
Hg content in the foliage was analyzed. Foliar Hg increased with leaf age for two to three months and
then stabilized at leaf concentrations near 150 ng/g. About 80% of the Hg found in above-ground biomass
was present in the leaves. The concentration of Hg in trees grown in the same mesocosms in containers of
low Hg soil (0.03 ± 0.01 jxg/g) exhibited foliar Hg concentrations that were similar to those of trees grown
in Hg-enriched soil. Almost all of the foliar Hg originated from the atmosphere. Clearly, plant foliage can
be a major sink for airborne Hg, which can subsequently enter the soil after litterfall (Ericksen et al.,
2003). However, this study did not determine the extent to which atmospheric Hg was dry-deposited on
the foliage, as opposed to gaseous uptake through the stomata.
9.4.4.3. Deposition to Soil
As with mosses, accumulation of heavy metals in surface soils provides a general reflection of the
spatial distribution of industrial pollution. In the study of Romic and Romic (2003), relationships were
found between urban activities and concentrations of metals in soils in developed areas surrounding
Zagreb, Croatia. Goodarzi et al. (2002) compared deposition estimated by moss bags to concentrations of
metals in A-horizon soils in the vicinity of a large smelter. Statistically significant correlations were
December 2008
9-97
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
observed between the moss bag deposition estimates and the soil metal concentrations for Cd, Pb, Zn, and
in some cases also Cu. These correlations suggested that atmospheric deposition of metals caused
elevated metal concentrations in the uppermineral horizon of these soils. No correlations were found for
Hg or As in this study.
Studies have also looked at metal accumulation in peat because of the tendency of most metals to
be immobilized through binding with organic matter. Steinnes et al. (2005) presented geographical
patterns of metal concentrations in surface peat throughout Norway that corresponded to pollution
sources, although the peat samples were collected in 1979. Zaccone et al. (2008) found that variations of
metal concentrations with depth in a single Swiss peat core corresponded with the depositional history
that would be expected from the industrial revolution, although Cs137 activity exhibited a distribution in
the profile that was not fully consistent with the Chernobyl nuclear reactor accident. A detailed study of
Finish peat showed that relationships between depth profiles of metal concentrations and deposition
history can match well for some metals at some sites, but not well for the same metals at other sites
(Rausch et al., 2005). They also found that Zn and Cd accumulation rates were independent of deposition
history at each of three study sites.
Metal deposition to soil is also a significant concern adjacent to roadways. The concentrations of
Cd, Ca, Cu, Pb, and Zn in road runoff were found to be significantly higher during winter in Sweden. This
seasonal pattern was attributed to the intense wearing of the pavement that occurred during winter due to
the use of studded tires in combination with chemical effects of deicing salts (Backstrom et al., 2003).
9.4.5. Components of Deposition
9.4.5.1. Trace Metals
Atmospheric deposition can be the primary source of some metals to some watersheds. Metal
inputs can include the primary crustal elements (Al, Ca, K. Fe, Mg, Si, Ti) and the primary anthropogenic
elements (Cu, Zn, Cd, Cr, Mn, Pb, V). The crustal elements are derived largely from weathering and
erosion, whereas the anthropogenic elements are derived from combustion, industrial sources, and other
man-made sources (Goforth and Christoforou, 2006).
There are a number of natural geologic sources of Hg emissions to the atmosphere. These include
geothermal and volcanic activity, geologic metal deposits, and organic-rich sedimentary rocks. These
natural emissions combine with anthropogenic emissions from such sources as power plants, landfills,
sewage sludge, mine waste, and incineration (Gustin, 2003; Schroeder and Munthe, 1998). Emissions
December 2008
9-98
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
from natural sources are controlled by geologic features, including substrate Hg content, rock type, the
degree of hydrothermal activity, and the presence of heat sources (Gustin, 2003).
The significance of natural Hg sources relative to anthropogenic sources varies geographically. For
example, Nevada occurs within a global mercuriferous belt, with area emissions about three times higher
than the value assumed for global modeling (Gustin, 2003). In Nevada, natural and anthropogenic Hg
emissions are approximately equal (Gustin, 2003).
Heavy metal deposition to forested sites depends on forest location as well as upwind emissions
source strength. The deposition velocity tends to be dependent on particle size and chemical species.
Larger particles deposit more efficiently than smaller particles. Heavy metals preferentially associate with
fine particles. Fine particles also have the longest atmospheric residence times. Depending on climate and
topography, fine particles may remain airborne for days to months and may be transported thousands of
kilometers from their source.
Ecosystems immediately downwind of major heavy metal emissions sources may receive locally
heavy dry deposition. Trace element investigations conducted in roadside, industrial, and urban
environments have also shown that substantial amounts of particulate heavy metals can accumulate on
vegetative surfaces.
The distribution of toxic elements in urban soils has been an important area of study (cfi, Madrid et
al., 2002; Markiewicz Patkowska et al., 2005). Generally, Cu, Pb, Zn, and Ni have accumulated in urban
soils compared with their rural counterparts (Yuangen et al., 2006). Effects on soil microbiology have not
been well studied but can include effects on microbial biomass, microbial utilization of C, and other
indicators of the health and functioning of urban soils. Yuangen et al. (2006) found that urban soil basal
respiration rates were positively correlated with soil acetic acid-extractable Cd, Cu, Ni, and Zn. The soil
microbial biomass was negatively correlated with the concentrations of Pb fractions, but not with other
metals. Overall microbial biomass was lower for urban soils as compared with rural soils (Yuangen et al.,
2006).
There is concern that Pb contamination of forest soil could move into groundwater. This would be
an important issue in view of the large quantity of Pb deposited from the atmosphere in the 1960s and
1970s in response to combustion of leaded gasoline. This issue was investigated by Watmough et al.
(2004) who applied a stable isotope (207Pb) to the forest floors of white pine (Pinus strobus) and sugar
maple (Acer saccharum) stands. Added Pb was rapidly lost from the forest floor, likely due to high litter
turnover in these forest types. However, Pb concentrations in the upper 30 cm of mineral soil were
strongly correlated with soil OM, suggesting that Pb does not readily move down the soil profile to the
ground water, but rather is associated with the organic content of the upper soil layers (Watmough et al.,
2004).
December 2008
9-99
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Karar et al. (2006) used factor analysis to identify possible sources contributing to PM10 deposition
at two urban sites in India: one residential and one industrial. At both sites, vehicular traffic and road dust
were identified as potential sources. In addition, solid waste dumping and soil dust were identified at the
residential site, and the galvanizing, electroplating, and tanning industries were identified at the industrial
site.
Azimi et al. (2003) found in rural and urban areas of France that dry deposition comprised 40, 60
and 80% of total deposition for Cd, Cu and Pb, although total concentrations were lower at the rural sites
than at the urban sites for most metals. Additional work by this group (Azimi et al., 2005) showed that
total deposition of these metals decreased from 1994 to 2002. Sabin et al. (2006a) presented literature
values of heavy metal dry deposition that showed urban areas to have deposition levels approximately an
order of magnitude higher than rural areas. Tasdamir et al. (2006) also reported deposition fluxes in dry
deposition for trace metals (Cu, Pb, Mn, Cr, Ni, Co, and Cd) in a Turkish urban environment. Fluxes were
of similar magnitude to those previously reported for urban industrialized areas, although the highest
particulate fluxes were for crustal elements (Ca, Mg, Fe and Mn) associated with coarse particles.
Information on atmospheric transport and deposition of heavy metals was provided by analysis of
snow chemistry radiating away from a metal smelter in an isolated region in Quebec (Telmer et al., 2004).
Transport of 27 metals was found to exceed 50 km. Wet deposition was distinguished from dry deposition
by filtering melted snow samples. Deposition beyond 50 km was found to be largely in soluble forms,
although significant particle deposition occurred beyond this distance. Partitioning between soluble and
insoluble forms varied by element. Elements that were most readily wet deposited included Pb and Cu,
and elements with the largest fraction in particles were Ti and Sb.
Detailed modeling results obtained by Lu et al. (2003) for the Los Angeles basin showed that the
majority (approximately 80%) of local metal deposition was associated with particles larger than 10 |_im.
and as a result, 35-45% of metal emissions were deposited locally. However, this study also indicated that
most of the remaining metals not deposited locally within the basin (65-75%) are transported over
continental to global scales. These authors further concluded that routine air monitoring for PM in the 2.5
and 10 mm size fractions is not adequate for measuring urban trace metal deposition. The importance of
large particle deposition was also demonstrated by Tasdemir et al. (2004) in Chicago, where the
deposition of polychlorinated biphenyls (PCBs) was associated with particles of about 25 (.un. Because
there are few point sources of PCBs, the gaseous phase of these molecules is expected to be in
equilibrium with PM-associated PCBs. The relatively high air concentrations and deposition velocities of
coarse particles in urban settings therefore lead to higher PCB deposition than in rural environments.
Much of the urban PMi0 emissions total is from non-exhaust traffic emissions (Hussein et al.,
2008). Studded tires cause higher PMi0 emissions than summer tires. Friction wear is small compared to
December 2008
9-100
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
suspension of accumulated road dust in contributing to the total PM10 emissions of traffic. Deposition of
heavy metals may be particularly high along freeways and other heavily traveled roads. Sabin et al.
(2006b) found that deposition rates in the vicinity of a large Los Angeles freeway were considerably
higher than urban background rates 10 to 150 m away from the freeway, particularly for Cu, Pb and Zn.
This was explained by a combination of vehicle emissions and resuspension of coarse particles (>6 |_im)
by rapidly moving traffic that provided PM to which the metals can adsorb. Similar results were observed
in Australia, where road-deposited sediments were found to have high concentrations of Zn, Fe, Pb, Cd,
Cu, Cr, Al and Mn (Herngren et al., 2006). However, this study found maximum metal concentrations
associated with particles in the 0.45 to 75 (.un range, which is somewhat smaller than those reported by
Sabin et al. (2006a). Herngren et al. (2006) also pointed out that the PM size class of 0.45 to 75 |_im was
much smaller than the minimum size of 250 |_im removed by street cleaning practices.
Because PM deposited along roadsides is prone to being washed into storm drains and ditches that
empty into nearby surface waters, the high deposition of trace metals associated with roads represents a
significant metal loading to surface waters (Herngren et al., 2006; Sabin et al., 2006a; 2006b) This has
also been shown in two studies of roof wash-off in urban areas. In Paris, France, Rocher et al. (2004)
found that wash-off from roofs was a significant input to surface waters for heavy metals and
hydrocarbons. They found that metal deposition did not have a strong seasonal signal, but that incomplete
combustion of heating fuels resulted in the highest hydrocarbon deposition during the heating season. A
similar study in Austin, TX showed high concentrations of Zn, Pb, Cd, and PAH in roof wash-off (Van
Metre and Mahler, 2003). In both the studies of Rocher et al. (2004) and Van Metre et al. (2003), roofing
materials had the largest effect on Zn concentrations in wash-off, but concentrations of Cu and Pb were
also elevated by roof materials.
9.4.5.2. Mercury
The most important factors involved in the atmospheric fate and transport of Hg include:
(1) emissions; (2) atmospheric transformation and transport; (3) deposition to the Earth surface; and
(4) re-emission to the atmosphere (U.S. EPA, 1997).
There are both anthropogenic and natural sources of Hg emission to the atmosphere. Natural
processes include volatilization from marine and fresh water aquatic ecosystems, degassing from soils and
geologic materials, and volcanic emissions (U.S. EPA, 1997). Most anthropogenic emissions are from
combustion sources and industrial processes. Particulate Hg emissions are mainly in oxidized form due to
the relatively high vapor pressure of elemental Hg (U.S. EPA, 1997).
The residence time of Hg in the atmosphere depends on its chemical form. Hg(0) has an average
atmospheric residence time of about a year, and is therefore transported long distances. Oxidized Hg
December 2008
9-101
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
(Hg(II)) has a residence time of hours to months and is prone to rapid deposition via wet and dry
deposition processes. However, some Hg(II) is associated with fine particles, and therefore can have an
atmospheric residence time that is more similar to that ofHg(O) (U.S. EPA, 1997). Fluxes of Hg(0) from
geologic sources, and anthropogenic emissions from combustion and industrial sources all contribute to
the global atmospheric reservoir of Hg, which has a residence time of up to a couple of years (U.S. EPA,
1997).
The divalent species of Hg are more readily removed from the atmospheric than is elemental Hg.
Dry deposition transfers both particulate and gaseous divalent Hg from the atmosphere to the Earth
surface at locations where substantial amounts of atmospheric divalent Hg occurs. In addition, particulate
Hg is readily wet deposited due to cloud scavenging processes and precipitation. The divalent species
have much lower Henry's Law constants than does elemental Hg and therefore partition readily to the
aqueous phase. This is especially true for the gas phase divalent Hg (U.S. EPA, 1997).
Most of the Hg emitted to the atmosphere deposits as Hg(II). The deposited Hg(II) can revolatilize
back to the atmosphere, be methylated in the soil, or be transported to a water body via runoff and
leaching. Methylation can also occur within the water body, and either Hg(II) or methyl Hg can be
reintroduced from the water back to the atmosphere.
Atmospheric Hg deposition has increased with industrialization. For example, Steinnes et al.
(2005) found that Hg concentrations in dated peat samples were about 15 times higher in the last 100
years than in pre-industrial times. However, the complexities of atmospheric Hg chemistry and typically
low atmospheric concentrations make quantification of particulate Hg deposition a challenging process
(Lynam and Keeler, 2005). Gaseous divalent Hg (Hg(II)) is highly soluble, and therefore can be
wet-deposited, but also can become associated with various types of PM that will control its deposition
characteristics. Gaseous elemental Hg (Hg(0)) is relatively insoluble, but can be oxidized in the
atmosphere to Hg(II) (Seigneur et al., 2003). Collection of PM Hg is typically done through air filtration
that requires extended sampling times as a result of low air concentrations. Lynam et al. (2005) found that
oxidation of Hg(0) may occur if sampling is done during high 03 events, resulting in artificially high
measurements of PM Hg concentrations. The complexities of Hg deposition were also shown by Graydon
et al. (2006) who found that a portion of wet deposited Hg(II) to forest canopies may be photo-reduced to
gaseous Hg(0) that is then reemitted to the atmosphere. This canopy effect suggests that throughfall may
underestimate total deposition. Net deposition of Hg to soil surfaces was evaluated by Gustin et al.
(2006). This study showed that reemission of deposited Hg from soil is affected by environmental
conditions including soil moisture, temperature, light, atmospheric oxidants and Hg concentrations in air.
Further work is needed to constrain estimates of deposition to soils under varying conditions.
December 2008
9-102
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Modeling approaches have been developed from assumptions regarding Hg deposition processes.
Without measurement data for evaluation, these results have an unknown level of uncertainty. However,
detailed Hg emissions inventories have been developed for some regions, such as the Great Lakes
(Murray and Holmes, 2004), which have been useful for deposition modeling. Using this approach,
Landis and Keeler (2002) estimated that dry deposition to Lake Michigan approximately equaled wet
deposition, and that atmospheric deposition to the lake was the primary Hg input. Several other studies
have developed deposition estimates for the Great Lakes through modeling. These include Gbor et al.
(2007), in which particulate Hg(II) was estimated to be only 7% of dry deposited gaseous Hg(II), and 2%
of total deposition to Lake Michigan.
Similar modeling approaches have been used in efforts to distinguish among deposition of locally
and regionally emitted Hg and the deposition of Hg that has resulted from intercontinental transport.
Seigneur et al. (2003) estimated that U.S. emissions other than from New York contributed 25 to 49% of
total Hg deposition in the U.S. This study also estimated that particulate Hg(II) comprised 1-2% of total
deposition in New York State, although dry deposition of gaseous Hg(II) was estimated to comprise about
30% of total Hg deposition.
Hg deposition measurements at Mace Head, Ireland, considered a global background site, were
reported by Seigneur et al. (2003) to be 14-94 pg/m3 for wet and gaseous deposition of Hg(II) and 5-115
pg/m3 for particulate Hg(II) deposition. A similar modeling exercise by Travnikov (2005) estimated that
for the Northern Hemisphere in general, the contribution of total Hg deposition from intercontinental
transport was about the same as that from regional pollution sources, although the Travnikov (2005)
article did not report the fraction of dry or particulate deposition. This literature suggests that particulate
Hg deposition is a small fraction of total Hg deposition, but the measurement difficulties and limited
available data coupled with the complexities of atmospheric Hg speciation indicate that further research is
needed to ascertain the environmental relevance of particulate Hg.
9.4.5.3. Organics
Most persistent organic pollutants (POPs) enter the biosphere via human activities, including
synthetic pesticide application, output of polychlorinated dibenzo dioxins (PCDD) from incinerators, and
accidental release of PCBs from transformers (Lee et al., 2003d). Once they are introduced into the
environment, their accumulation and magnification in biological systems are determined by
physiochemical properties and environmental conditions. Uptake by plants can occur at the soil/plant
interface and at the air/plant interface. For lipophilic POPs, such as PCDDs and PCBs, the air/plant
response route generally dominates (Lee et al., 2003d; Thomas et al., 1998), but uptake through
December 2008
9-103
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
above-ground plant tissue also occurs. In a study of zucchini (Cucurbita pepo), Lee et al. (2003d) found
chlordane pesticide components in all vegetation tissues examined: root, stem, leaves, fruits.
Organic compounds partition between gas and particle phases, and organic particulate deposition
depends largely on the particle sizes available for adsorption (U.S. EPA, 2004). Dry deposition of organic
materials is often dominated by the coarse fraction. Gas-particle phase interconversions are important in
determining the amount of dry deposition.
Many pesticides are carcinogenic or estrogenic and pose potential threats to aquatic and terrestrial
biota. Although deposition of semi-volatile organic compounds (SOC) was previously reported for the
Sierra Nevada Mountains in California and the Rocky Mountains in Colorado, little was previously
known about the occurrence, distribution, or sources of SOC in alpine, sub-Arctic, and Arctic ecosystems
in the western U.S. The snowpack is efficient at scavenging of both particulate and gas phase pesticides
from the atmosphere (Halsall, 2004; Lei and Wania, 2004).
Analysis of pesticides in snowpack samples from seven national parks in the western U.S. by
Hageman et al. (2006) illustrated the deposition and fate of 47 pesticides and their degradation products.
Correlation analysis with latitude, temperature, elevation, PM, and two indicators of regional pesticide
use suggested that regional patterns in historic and current agricultural practices are largely responsible
for the distribution of pesticides in the national parks. Pesticide deposition to parks in Alaska was
attributed to long-range atmospheric transport.
Deposition and fate of PAH has been an important area of research. Because they are carcinogenic,
PAHs are important environmental contaminants. Root-soil behavior of PAHs is an area of active study.
Soil-bound PAHs are associated with soil organic matter and are therefore generally not easily available
for root uptake. PAHs are readily adsorbed to root surfaces but there seems to be little movement to the
interior of the root or movement up to the shoots (Gao and Zhu, 2004).
Paddy rice is the main food crop planted in China. As an aquatic plant having aerial roots, the
movement of PAHs into rice roots may be different than their movement into more widely studied
land-grown food crops. PAH concentrations in the rice roots were more correlated with the water and air
compartments than with the soil (Jiao et al., 2007).
The group PAH contaminant group includes known carcinogens, such as benzo[a]pyrene (B[a]P)
and substances thought to be toxic. They are common air pollutants in metropolitan areas, derived from
vehicular traffic and other urban sources. Especially high concentrations have been found near Soderberg
aluminum production industries and areas where wood heating during winter is common. Other sources,
in addition to gasoline and diesel engines, include forest fires and various forms of fossil fuel combustion
(Sanderson and Farant, 2004).
December 2008
9-104
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
PAHs include hundreds of different compounds that are characterized by possessing two or more
fused benzene rings. They are widespread contaminants in the environment, and are formed by
incomplete combustion of fossil fuels and other organic materials. Eight PAHs are considered
carcinogenic and 16 are classified by EPA as priority pollutants. The behavior of PAHs is strongly
determined by their chemical characteristics, especially their nonpolarity and hydrophobicity. They
readily adsorb to particulates in the air and to sediments in water. Srogi (2007e) provided a thorough
review of PAH concentrations in various environmental compartments and their use for assessing
environmental risks and possible effects on ecosystems and human health.
The total PAH concentration in grasses adjacent to a highway have been measured to be about eight
times higher than in grasses from reference sites not close to a highway (Crepineau et al., 2003). Howe et
al. (2004) found that concentrations of PAHs and hexachlorobenzene (HCB) in spruce (Picea spp.)
needles at 36 sites in eastern Alaska varied by an order of magnitude. Samples collected near the city of
Fairbanks generally had higher concentrations than samples collected from rural areas. The relative
importance of combustion sources versus petrogenic sources was highest in the near-coastal areas, as
reflected in variation in the concentration of ratios of isomeric PAHs.
9.4.5.4. Base Cations
With respect to ecosystem effects from PM deposition, the inclusion of base cations (especially Ca,
Mg, and K) in atmospheric deposition is generally considered to be a positive effect. Base cations are
important plant nutrients that are in some locations present in short supply and that are further depleted by
the acidic components of deposition. Increased base cation deposition can help to ameliorate adverse
effects of acidification of soils and surface waters and reduce the toxicity of inorganic Al to plant roots
and aquatic biota. These topics are covered in detail in the recent NOxSOx ISA (2008c).
Calcium supply is also well known to be important for breeding success in passerine bird species.
Eggshell thickness, egg size, clutch size, and hatchability of pied flycatcher (Ficedula hypoleuca) were
found to be depressed near the Cu smelter at Haijavalta, SW Finland (Eeva and Lehikoinen, 2004).
Availability of Ca-rich food to the birds was estimated by counting snail shells in the nests postfledging.
The number of snail shells correlated positively with the Ca concentration of nestling feces and adult
breeding success. In addition, the negative impact of Cu on the number of fledglings was stronger at
locations where Ca concentration was low (Eeva and Lehikoinen, 2004).
Although the effects of base cation deposition inputs to terrestrial ecosystems are most commonly
considered to be positive, under very high base cation deposition levels, plant health can be adversely
affected. Dust that is high in base cations can settle on leaves and other plant structures and remain for
extended periods of time. This is especially likely in arid environments because rainfall can serve to wash
December 2008
9-105
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
dry deposited materials off the foliage. Extended dust coverage can result in a variety of adverse impacts
on plant physiology (Grantz et al., 2003). For example, van Heerden et al. (2007) documented decreased
chlorophyll content, inhibition of C02 assimilation, and uncoupling of the oxygen-evolving complex in
desert shrubs exposed to high limestone dust deposition near a limestone quarry in Namibia.
Based on the IFS data, the U.S. EPA (2004) concluded that particulate deposition has a greater
effect on base cation inputs to soils than on base cation losses associated with the inputs of sulfur,
nitrogen, and H . These atmospheric inputs of base cations have considerable significance, not only to the
base cation status of these ecosystems, but also to the potential of incoming precipitation to acidify or
alkalize the soils in these ecosystems. This topic is discussed in detail in the recent NOxSOx ISA (2008e).
9.5. Effects on Individual Organisms
Deposition of PM from the atmosphere to the soil or plant surface is required before most
biological effects on plants or ecosystems can occur. Exposure to a given amount of airborne PM may
lead to differing responses, depending on the particular mix of deposited particles. PM is not a single
pollutant, but rather a heterogeneous mixture of particles differing in size, origin, and chemical
composition. Atmospheric PM has been defined, for regulatory purposes, mainly by size fractions and
less clearly so in terms of chemical nature, structure, or source. PM size classes do not necessarily relate
to effects (U.S. EPA, 1996). Both fine and coarse-mode particles may affect plants and other organisms.
Much of the burden of sulfates (S042 ). nitrates (N03 ). ammonium salts (NH/), and hydrogen ions (H )
resides in the atmosphere either dissolved in fog water or as liquid or solid aerosols. Assessment of
atmospheric deposition effects of S and N particles overlaps substantially with material covered in the
recent NOxSOx ISA (2008e). Therefore, effects of acidifying particulate deposition are not covered in this
assessment.
9.5.1. Effects on Plants
Exposure to airborne PM can lead to differing phytotoxic responses, depending on the particular
mix of deposited particles. This was well-known at the time of the previous PM criteria assessment, as
summarized below. Effects of particulate deposition on individual plants or ecosystems are difficult to
characterize because of the complex interactions among biological, physicochemical, and climatic factors.
Most direct effects occur in severely polluted areas surrounding industrial point sources, such as
limestone quarries, cement kilns, and metal smelting facilities (U.S. EPA, 2006b). Experimental
application of PM constituents to foliage typically elicits little response at the more common ambient
December 2008
9-106
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
exposure concentrations. The diverse chemistry and size characteristics of ambient PM and the lack of
clear distinction between effects attributed to phytotoxic particles and to other air pollutants further
confound understanding of the direct effects on foliar surfaces. The majority of the documented toxic
effects of particles on vegetation reflect their chemical content (e.g., acid/base, trace metal, nutrient),
surface properties, or salinity (U.S. EPA, 2004).
Studies of the direct effects of particles on vegetation have not yet advanced to the stage of
reproducible exposure experiments. In general, phytotoxic gases are deposited more readily, assimilated
more rapidly, and lead to greater direct injury of vegetation as compared with most common particulate
materials. The dose-response functions obtained in early experiments following the exposure of plants to
phytotoxic gases generally have not been observed following the application of particles (U.S. EPA,
2004).
Atmospheric PM may affect vegetation directly following deposition on foliar surfaces or
indirectly by changing the soil chemistry or by changing the amount of radiation reaching the Earth's
surface through PM-induced climate change processes. Indirect effects acting through the soil are often
thought to be most significant because they can alter nutrient cycling and inhibit plant nutrient uptake
(U.S. EPA, 2004).
Particles can be deposited from the atmosphere to surfaces of the leaf, twig, or bark. Subsequently,
those particles can be taken up by the plant through the leaf surface, or be removed from the plant via
resuspension to the atmosphere, washing by rainfall, or litter-fall with subsequent transfer to the soil. Any
PM deposited on above-ground plant parts can have physical or chemical effects on the plant. The U.S.
EPA (2004) reported that the effects of "inert" PM are mainly physical; whereas those of toxic particles
can be both chemical and physical.
Since publication of EPA's 2004 PM criteria assessment, additional research has been conducted on
the effects of PM on plants. For example, windblown PM affects physical, chemical, and biological
attributes of both plants and animals (c.f., Englert, 2004; Gleason et al., 2007; Kappos et al., 2004).
Experiments by Gleason et al. (2007) suggest that most direct effects on plants of windblown PM
originating on road surfaces occur within 40 m of the source. Windblown PM from roads or agriculture
can cover plant photosynthetic structures (Sharifi et al., 1999), cause impact damage (Armbrust and Retta,
2002), or interfere with physiological mechanisms (Burkhardt et al., 2002).
The atmospheric deposition of PM into the ocean has important implications for primary
productivity and carbon sequestration. This is because metals in PM deposition limit phytoplankton
growth in parts of the ocean (Crawford et al., 2003). In particular, Fe and Zn can influence the
productivity of algae that are involved in CaC03 production. The production of both particulate organic C
and CaC03 drive the ocean's biological carbon pump (Shulz et al., 2004). Thus, in oceanic areas of trace
December 2008
9-107
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
metal limitation, changes in trace metal atmospheric deposition can affect biogenic calcification, with
potential consequences for C02 partitioning between the ocean and atmosphere.
A study by Sheesley et al. (2004) illustrated the value of bioassay procedures to provide an initial
screening of ambient PM toxicity. They used two species of green algae and two extraction methods to
compare the toxicities of atmospheric PM collected at two urban/industrial sites and one rural site near the
southern shore of Lake Michigan. Toxicities varied by site, by extraction solvent, and by bioassay. Results
suggested that toxicity was not related to the total mass of PM in the extract, but to the chemical
components of the PM. It is noteworthy that the concentrations of PAHs and other contaminants in PM in
this type of short-term and acute toxicity testing are much higher than would be found in the natural
environment. Thus, the purpose of this type of testing is to provide an initial screening-level comparison
of relative toxicities of atmospheric PM from different source areas. It does not provide the data that
would be needed to assess risk (Sheesley et al., 2004).
Some plant species have good ability to extract heavy metals from soil, thereby offering potential
for phytoremediation. For example, several species of willow (Salix spp.) accumulate high levels of Zn
and Cd in aboveground biomass (Lunackova et al., 2003; Meers et al., 2007; Rosselli et al., 2003). A first
estimation of the order of magnitude of potential metal removal by willow was 2 to 27 kg/ha/yr of Zn and
0.25 to 0.65 kg/ha/yr for Cd (Meers et al., 2007).
Otnyukova (2007) demonstrated vertical gradients within a coniferous forest canopy in the
fruticose lichen genus Usnea with respect to lichen thallus morphology and heavy metal concentration.
Abnormal thalli at the tree-top level contained higher concentrations of Al, Fe, Zn, F, Sr, and Pb. This
vertical pattern within the tree canopy is in general accordance with known deposition of PM to plants
(Otnyukova, 2007).
9.5.1.1. Direct Effects of Coarse-mode Particles
The current state-of-scientific knowledge regarding the direct effects of coarse PM on plants has
not changed since publication of the previous PM criteria assessment (U.S. EPA, 2004). The summary
provided here is taken from that report. In many rural areas and some urban areas, the majority of the
mass in the coarse particle mode derives from the elements silicon, aluminum, calcium, and iron,
suggesting a crustal origin as fugitive dust from disturbed land, roadways, agriculture tillage, or
construction activities. Rapid sedimentation of coarse particles tends to restrict their direct effects on
vegetation largely to roadsides and forest edges, which often receive the greatest deposition (U.S. EPA,
2004).
Dust can cause both physical and chemical effects. Deposition of inert PM on above-ground plant
organs sufficient to coat them with a layer of dust may result in changes in radiation received, a rise in
December 2008
9-108
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
leaf temperature, and the blockage of stomata. Crust formation can reduce photosynthesis and the
formation of carbohydrates needed for normal growth, induce premature leaf-fall, damage leaf tissues,
inhibit growth of new tissue, and reduce starch storage. Dust may decrease photosynthesis, respiration,
and transpiration; and it may result in the condensation and reactivity of gaseous pollutants with PM,
thereby causing visible injury symptoms and decreased productivity (U.S. EPA, 2004).
The chemical composition of PM is usually the key phytotoxic factor leading to plant injury. For
example, cement-kiln dust liberates calcium hydroxide on hydration. It can then penetrate the epidermis
and enter the mesophyll, causing an increase in leaf surface pH.
Sea-salt particles can serve as nuclei for the absorption and subsequent reaction of other gaseous
and particulate air pollutants. Direct effects on vegetation reflect these inputs and salt injury caused by the
sodium and chloride that constitute the bulk of these particles. Foliar injury from salt deposition can
influence plant species composition in coastal environments. It appears that, to cause injury, salt deposited
on leaf surfaces must dissolve and be absorbed into leaf tissue. Therefore, if RH remains below about
70%, even heavy deposition of salt may not induce injury to some plant species (U.S. EPA, 2004).
Little salt is taken up by plant roots; rather, most enters through the aerial organs. Mechanical
injury resulting from leaves and twigs beating against each another in the wind at coastal locations causes
the formation of small lesions through which salt can enter. After entry into the plant, chloride can be
translocated to the leaves and twigs where it can accumulate to concentrations that kill a portion of the
plant. Deposition and translocation of chloride results in death of the seaward leaves and twigs. The result
is the continued growth of the uninjured branches in an inland direction. As a result, the canopy angle
varies with the intensity of salt spray (U.S. EPA, 2004).
Injury to vegetation from the application of deicing salt is caused by salt spray blown or drifting
from the highways (Viskari and Karenlampi, 2000). The most severe injury is often observed nearest the
highway. Conifers planted near roadway margins in the eastern U.S. often exhibit foliar injury due to
toxic levels of saline aerosols deposited from deicing solutions (U.S. EPA, 2004).
9.5.1.2. Effects of Fine-mode Particles
Trace Elements
Effects of fine particle trace elements were described by the U.S. EPA (2004), and some additional
more recent research has also been conducted, especially on the topic of vegetative uptake of trace
elements from the soil. The state of scientific understanding as presented by the U.S. EPA (2004) is
summarized below, followed by discussion of more recent research findings. All but 10 of the 90 elements
that comprise the inorganic fraction of the soil occur at concentrations of < 0.1% (1000 jxg/g) and are
December 2008
9-109
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
termed "trace" elements or trace metals. Trace metals with a density greater than 6 g/cm3, referred to as
"heavy metals," are of particular interest because of their potential toxicity to plants and animals.
Although some trace metals are essential for vegetative growth or animal health, they are all toxic in large
quantities. Most trace elements exist in the atmosphere in particulate form as metal oxides (Ormrod,
1984). Aerosols containing trace elements derive predominantly from industrial activities. Generally, only
the heavy metals Cd, Cr, Ni, and Hg are released from stacks in the vapor phase (McGowan et al., 1993).
Deposition of trace elements along roadsides and in industrial and urban environments can cause
accumulation of particulate heavy metals on vegetative surfaces. Foliar uptake of metals can cause
adverse effects in aboveground plant tissues. Low solubility limits foliar uptake and direct heavy metal
toxicity because trace metals must be brought into solution before they can enter into the leaves or bark of
vascular plants. In those instances when trace metals are absorbed, they are frequently bound in leaf tissue
and are lost when the leaf drops off (Hughes, 1981).
Direct effects of trace elements on vegetation can result from their deposition and residence on
foliar surfaces. Low solubility limits entry into plant tissue. Trace metals absorbed into leaf tissue are
eventually transferred to the soil litter layer where they can affect litter decomposition, an important
source of soil nutrients. Fungi and microorganisms living on leaves aid in leaf decomposition after leaves
are dropped to the forest floor. Changes in litter decomposition processes in response to metal toxicity can
influence nutrient cycling in the soil and limit the supply of essential nutrients.
Trace metals, particularly heavy metals (e.g., Cd, Cu, Pb, Cr, Hg, Ni, Zn) can influence forest
growth. Growth suppression of foliar microflora has been shown to result from Fe, Al, and Zn. These
three metals can also inhibit fungal spore formation, as can Cd, Cr, Mg, and Ni (see Smith, 1990). The
greatest injury to vegetation occurs from pollution near mining, smelting, and other industrial sources.
Direct metal phytotoxicity can occur only if the metal can move from the surface into the leaf or directly
from the soil into the root.
Trace metals are found in small amounts in soils, ground water, and vegetation. Many are essential
micronutrients required for growth by plants and animals. Naturally occurring mineralization can produce
metal concentrations in soils and vegetation that are high compared to atmospheric sources. Many metals
are bound by chemical processes in the soil, reducing their availability to biota. However, epiphytic or
parasitic root colonizing microorganisms can solubilize and transport metals for toot uptake (cf., Lingua
et al., 2008).
Heavy metals deposited from the atmosphere to forests accumulate either in the organic forest floor
or in the upper mineral soil layers. These are the areas that have the greatest root development. Metal
concentration tends to decrease with soil depth. Shallow-rooted plant species are most likely to take up
metals from the soil (Martin and Coughtrey, 1981). Though all heavy metals can be directly toxic at
December 2008
9-110
DRAFT-DO NOT QUOTE OR CITE
-------
1 sufficiently high levels, only Cu, Ni, and Zn have been documented as frequently being toxic to plants
2 (U.S. EPA, 2004). Toxicity due to Cd, Co, and Pb has been seen only under unusual conditions (Smith,
3 1990). Chronic exposure at lower concentrations have the potential to interfere with nutrient-cycling
4 processes if mycorrhizal function is impaired. The potential pathways of accumulation of trace metals in
5 terrestrial ecosystems, as well as the possible consequences of trace metal deposition on ecosystem
6 functions, are summarized in Figure 9-53 (U.S. EPA, 2004).
Stem Flow
¦¦ !¦ ¦¦ ition
Biologically
Available,
Biologically
Unavailable
10. Root
V Turnover
5. Mass Flow,
_ Diffusion
11. Mineralization
IX.
Soil Organic
Upper Soil
VI.
Root Storage,
Metabolism
7. Leaching
6. Root
Uptake
12. Weathering
VII.
Lower Soil
Primary
Minerals
5. Mass Flow,
Diffusion
^ 7. Leaching
VIII.
Groundwater
Source: U.S. EPA, (2004)
Figure 9-53. Relationship of plant nutrients and trace metals with vegetation. Compartments
(roman numerals) represent potential storage sites; whereas arrows (Arabic
numerals) represent potential transfer routes.
7 The effects of Pb on ecosystems are discussed in the 2006 Pb AQCD (U.S. EPA, 2006b), which
8 concluded that, due to the deposition of Pb from past practices (e.g., leaded gasoline, ore smelting) and
9 the long residence time of Pb in many aquatic and terrestrial ecosystems, a legacy of environmental Pb
10 burden exists, over which is superimposed much lower contemporary Pb loadings. The potential for
December 2008
9-111
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
ecological effects of the combined legacy and contemporary Pb burden to occur is a function of the
bioavailability or bioaccessibility of the Pb, which, in turn, is highly dependent upon numerous site
factors (e.g., soil organic carbon content, pH, water hardness). However, while the more localized
ecosystem impacts observed around smelters are often striking, it was found that the effects could not be
attributed solely to Pb, recognizing the presence of many other stressors (e.g., other heavy metals, oxides
of sulfur and nitrogen) that can also act singly or in concert with Pb to cause such notable environmental
impacts. (U.S. EPA, 2004, 2008e).
At the time of the most recent air quality criteria report for PM (U.S. EPA, 2004), trace metal
toxicity of lichens had been demonstrated in relatively few cases. Nash (1975) documented Zn toxicity in
the vicinity of a Zn smelter near Palmerton, PA. Experimental data had suggested that lichen tolerance to
Zn and Cd generally ranges between 200 and 600 ppm (Nash, 1975).
Phytochelatins are intracellular metal-binding peptides that act as specific indicators of metal
stress. Because they are produced by plants as a response to sublethal concentrations of heavy metals,
they can indicate that heavy metals play a role in forest decline (Gawel et al., 1996). Phytochelatin
concentrations have previously been measured in coniferous trees in the northeastern U.S. The U.S. EPA
(2004) summarized studies indicating that both the number of dead red spruce trees and phytochelatin
concentrations increased sharply with elevation in the northeastern U.S. Red spruce stands showing
varying degrees of decline indicated a systematic and significant increase in phytochelatin concentrations
associated with the extent of tree injury. The U.S. EPA (2004) interpreted these data as indicating that
metal stress causes tree injury and contributes to forest decline in the northeastern U.S.
Mercury in vegetation is derived almost exclusively from the atmosphere (Grigal, 2003). Mercury
uptake from soil is limited, partly because roots adsorb Hg but transport it to foliage very poorly (Grigal,
2002). Grigal (2003a) provided a thorough review of the sequestration of Hg in forest and peatland
ecosystems. A fundamental aspect of Hg cycling is its strong relationship to organic matter. For that
reason, peatlands sequester much larger quantities of Hg than would be expected on the basis of their land
area. Thus, if global climate change affects C storage, it may indirectly affect Hg storage because of the
strong relationship between Hg and organic matter (Grigal, 2003).
The accumulation of heavy metals in soils is determined by a variety of soil characteristics,
including pH, Fe and Al oxide content, amount of clay and organic material, and CEC (Hernandez et al.,
2003). Thus, the pattern of distribution of heavy metals in soils depends on the soil characteristics and on
the metal characteristics.
Small roots (< 2 mm diameter) provide the major uptake and transport system to the above-ground
plant and generally contain a large proportion of the total metals found in plants (Gordon and Jackson,
2000). Atmospherically-deposited metals accumulate in upper soil horizons where fine roots are most
December 2008
9-112
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
developed. Surface litter decomposition is reduced in soils having high metal concentrations. This is
likely due to the sensitivity to metals of microbial decomposers and reduced palatability of plant litter
having high metal concentration (Johnson and Hale, 2008). Root decomposition is a key component of
nutrient cycling. Johnson and Hale (2008) measured in situ fine root decomposition at Sudbury, Ontario
and Rouyn-Noranda, Quebec. Elevated soil metal concentrations (Cu, Ni, Pb, Zn) did not necessarily
reduce fine root decomposition. Only at sites having high concentrations of metals did decomposing roots
show increased metal concentrations over time.
The availability for plant uptake of metals in soil depends on metal speciation and soil pH. In
addition, metal binding to dissolved organic matter (DOM) reduces bioavailability (Sauve, 2001).
Because OM typically decreases with soil depth, the affinity of metals for OM can influence metal
bioavailability at different soil depths.
Heavy metal particles are important constituents of tire dust. These particles accumulate on the
road surface as part of brake linings, road paint, tire debris, DE, road construction materials, and catalyst
materials. Tire dust can be suspended in the atmosphere and contribute metals to soil, air, and urban
runoff (Adachi and Tainosho, 2004; Davis et al., 2001; Smolders and Degryse, 2002). In particular, Zn
oxide comprises 0.4 to 4.3% of tire tread (Smolders and Degryse, 2002) and tire dust is a substantial
source of environmental Zn pollution. Adachi and Tainosho (2004) used a field emission screening
electron microscope equipped with an energy dispersive x-ray spectrometer to characterize heavy metal
particles embedded in tire dust. Samples were classified into four likely source categories, based on
cluster analysis. Based on morphology and chemical composition, the samples were identified as having
derived from brake dust (rich in Fe, with traces of Cu, Sb, Ba), yellow paint (CrPb04 particles), brake
dust (particulate Ti, Fe, Cu, Sb, Zr, Ba and heavy minerals [Y, Zr, La, Ce]), and tire tread (Zn oxide).
Build-up of high concentrations of trace metals in soil is difficult to remediate because of the long
residence times of metals in the environment. Plants that survive on heavy metal contaminated soils are of
particular interest because of the mechanisms that allow them to tolerate such conditions and interactions
between soil contamination and vegetation composition (Becker and Brandel, 2007; Hall et al., 2002).
Burt et al. (2003b) investigated the concentrations and chemical forms of trace metals in
smelter-contaminated soils collected in the Anaconda and Deer Lodge Valley area of Montana, one of the
major mining districts of the world for over a century (1864 to 1983). The relative distributions of trace
metals within the more soluble soil extraction forms were similar to their respective total concentrations.
This suggested a relationship between the concentrations of total trace elements and concentrations of
soluble mobile fractions.
Sequential extractions do not provide direct characterization of trace metal speciation, but rather an
indication of chemical reactivity (Burt et al., 2003b; Ramos et al., 1994). Soluble and exchangeable forms
December 2008
9-113
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
are considered readily mobile and bioavailable. Those bound to clay minerals or organic matter are
considered generally unavailable.
There are substantial differences among plant species in their response to heavy metal exposure.
These differences can be attributed to differential uptake and excretion rates, increased storage capability,
and various physiological changes to compensate for metal stress. Toxicity response is also dependent on
the nutritional status of the plant and the development of mycorrhizae (Strandberg et al., 2006).
Under high atmospheric pollution levels, the abundance of most plant species tends to decrease
with increasing heavy metal concentrations in plant tissues. Salemaa et al. (2004) investigated heavy
metal concentrations in understory plant species growing at varying distances from the Harjavalta Cu-Ni
smelter. Heavy metal concentrations (except Mn) were highest in bryophytes, followed by lichens, and
were lowest in vascular plants. Vascular plants are generally able to restrict the uptake of toxic elements,
and therefore grew closer to the smelter than lichens. A pioneer moss (Pohlia nutans) was unusual in that
it survived close to the smelter despite its accumulation of high levels of Cu and Ni.
Contamination of stream sediments by heavy metals can impact adjacent terrestrial ecosystems
when high flows cause resuspension of sediment particles. For example, Ozdilek et al. (2007) showed that
metal concentrations in vegetation along the Blackstone River in Massachusetts and Rhode Island were
generally inversely related to the distance from the riverbank, with higher metal concentrations in plant
tissues located near the river.
Plants respond to high concentrations of metals in soil through a variety of mechanisms. These can
include exclusion, adaptation, compartmentalization, and chelation with phytochelatins, which are
peptides synthesized from glutathione.
It can be difficult to assess the extent to which observed metal concentrations in soil are of
anthropogenic origin. This is because soil parent material, pedogenesis, and anthropogenic inputs all
influence the amounts and distribution of trace elements in soil. Trace element concentrations in some
natural soils that are remote from air pollution can be higher than soils derived from other parent material
that receive anthropogenic inputs (Burt et al., 2003a).
Organic Compounds
Volatile organic compounds in the atmosphere are partitioned between the gas and particle phases.
As described by the U.S. EPA (2004), the partitioning depends on vapor pressure, temperature, surface
area of the particles, and the nature of the particles and of the chemical being adsorbed. A wide variety of
organic contaminants are deposited from the atmosphere. These include chemicals such as DDT, PCBs,
and PAHs.
December 2008
9-114
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Below is a summary of the findings of the U.S. EPA (2004), followed by discussion of more recent
research findings. Plants may be used as passive monitors to compare the deposition of organic
compounds between sites. Vegetation can be used semi-quantitatively to indicate organic pollutant levels
if the mechanism of accumulation is considered. Organic compounds can enter the plant via the roots or
be deposited as a particle onto the leaves and be taken up through the cuticle or stomata. The pathways
depend on the chemical and physical properties of the pollutant. These include, for example, lipophilicity,
water solubility, vapor pressure, and Henry's law constant. Environmental conditions can also be
important, including temperature and organic content of soil; plant species, and the foliar surface area and
lipid content.
Organic particulates in the atmosphere are diverse in their makeup and sources. Vegetation itself is
an important source of hydrocarbon aerosols. Terpenes, particularly a-pinene, (3-pinene, and limonene,
released from tree foliage may react in the atmosphere to form submicron particles. These naturally
generated organic particles contribute significantly to the blue haze aerosols formed naturally over
forested areas (Geron et al., 2000; U.S. EPA, 2004). The low water solubility with high lipo-affinity of
many organic xenobiotics control their interaction with the vegetative components of natural ecosystems.
Foliar surfaces are covered with a waxy cuticle layer that helps reduce moisture loss and short-wave
radiation stress. This epicuticular wax consists largely of long-chain esters, polyesters, and paraffins.
These accumulate lipophilic compounds. Organic air contaminants in the particulate or vapor phase can
be absorbed to, and accumulate in, the epicuticular wax of leaf surfaces. Direct uptake of organic
contaminants through the cuticle and the vapor-phase uptake through the stomata are not well
characterized for most trace organics.
The leaves of Quercus ilex have been shown to readily accumulate PAHs in situ. Young leaves
accumulated PAHs within three weeks of bud break. Mature leaves showed seasonality, with higher PAH
concentrations during winter (Alfani et al., 2005). Plants also vary in the extent to which they take up
heavy metals from the soil. Variability has been shown to occur in response to different plant species and
different metals. For example, Szabo and Fodor (2006) exposed winter wheat (Triticum aestivum), maize
(Zea mays) and sunflower (Helianthus annuus) to a variety of micro-pollutants. Cadmium accumulation
was significant in both vegetative and reproductive plant parts. Vegetative winter wheat accumulated
substantial amounts of Hg, but the other species did not. Lead, Cu, and Zn showed only moderate
enrichment in crops (Szabo and Fodor, 2006).
Topographic and vegetative characteristics exert different influence on deposition modes. In
general, dry deposition is most affected by plant morphology (Grantz et al., 2003). The potential effects of
PM on vegetation include the full range of biological organization, with exposures occurring through the
soil and through vegetative surfaces. In general, soil-mediated exposure is thought to be more significant
December 2008
9-115
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
(Grantz et al., 2003). Soil acts as an important storage compartment for POPs, including PCBs and PAHs.
There is a continuous process of partitioning between the soil pool and the atmosphere, and this controls
the regional and global transport of these compounds (Backe et al., 2004; Wania and Mackay, 1993). Over
time, POPs move towards equilibrium between the environmental compartments, and this process can be
described using the fugacity concept (Backe et al., 2004; Mackay, 1991). Fugacity reflects the tendency of
a chemical constituent to escape one environmental compartment and move to another. When an
equilibrium distribution is achieved, the fugacity quotient values in each compartment will be equal.
Soil/air partitioning is controlled by a variety of factors. These include soil properties, such as OM
content, moisture, porosity, texture, and structure, as well as the physiochemical properties of the
pollutant, including vapor pressure and water solubility.
9.5.2. Effects on Animals
Some amphibian ecotoxicological research has focused on heavy metal exposure. Contaminant
uptake can occur by oral, pulmonary, and dermal exposure (c.f., James et al., 2004; Johnson et al., 1999;
Lambert, 1997). This is potentially important because of documented declines in amphibian populations
in the U.S. and elsewhere in recent decades (c.f., Houlahan et al., 2000). Toads were shown to be fairly
tolerant of Cd exposure (James et al., 2004). It is not clear whether current levels of terrestrial metal
contamination pose an increased risk to amphibians in general.
Bioindicator organisms can be especially useful for monitoring PM effects over geographical and
temporal scales. Terrestrial invertebrates have been used to monitor contaminants in both air and soil.
Snails (Helix spp.) accumulate trace metals and agrochemicals, and can be used as effective biomonitors
for urban air pollution (Beeby and Richmond, 2002; Regoli et al., 2006; Viard et al., 2004). Demonstrated
biological effects include growth inhibition, impairment of reproduction, and induction of
metallothioneins that are involved in metal detoxification (Gomot-de Vaufleury and Kerhoas, 2000;
Regoli et al., 2006). The use of sentinel species to detect the effects of complex mixtures of air pollutants
is of particular value because the chemical constituents are difficult to characterize, exhibit varying
bioavailability, and are subject to various synergistic effects.
Regoli et al. (2006) caged land snails (Helix aspersa) at five locations in the urban areas of Ancona,
Italy. After four weeks of exposure to ambient air pollution levels, the snails were analyzed for trace
metals and PAHs. Biomarkers were measured that correlated with contaminant accumulation, including
levels of metallothioneins, activity of biotransformation enzymes, and peroxisomal proliferation. In
addition, indicators of oxidative stress were measured, such as oxyradical scavenging capacity, onset of
cellular damage, and loss of DNA integrity. Results documented substantial accumulation of metals and
December 2008
9-116
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
PAHs in snail digestive tissues in urban areas having high traffic congestion. Cellular reactivity was also
found, suggesting that this species is an effective bioindicator for multipollutant air quality and PM
monitoring.
Earthworms are considered to be relatively sensitive indicators of soil metal contamination. They
are continuously exposed to the soil via dermal contact and also ingest large quantities of soil. In addition,
earthworms often constitute a large percentage of soil animal biomass. Massicotte et al. (2003) compared
the cell viability and phagocytic potential of three earthworm species (Lumbricus terrestris, Eisenia
andrei, and Aporrectodea tuberculata) in response to atmospheric emissions of metals from a cement
factory in Quebec, Canada. Cell viability actually increased in proximity (0.5 km) to the cement factory
for A tuberculata, and this might have been due to beneficial effects of increased Ca deposition. There
were no significant differences observed for the other two species (Massicotte et al., 2003).
9.5.3. Effects on Microbes and Fungi
Accumulation of heavy metals in litter can interfere with nutrient cycling. Microorganisms are
responsible for decomposition of organic matter, which contributes to soil fertility. Toxic effects on the
microflora can be caused by Zn, Cd, and Cu. The U.S. EPA (2004) judged that addition of only a few
mg/kg of soil of Zn can inhibit sensitive microbial processes. Enzymes involved in the cycling of N, P,
and S (especially arylsulfatase and phosphatase) seem to be most affected (Kandeler et al., 1996).
It is believed that increased accumulation of litter in metal-contaminated areas is due to the effects
of metal toxicity on microorganisms. Smith (1991) reported the effects of Cd, Cu, Ni, and Zn on the
symbiotic activity of fungi, bacteria, and actinomycetes. In particular, the formation of mycorrhizae has
been shown to be reduced when Zn, Cu, Ni, and Cd were added to the soil.
Most studies of the effects of heavy metals on soils have been conducted under laboratory
conditions. However, Oliveira and Pampulha (2006) performed a field study to evaluate long-term
changes in soil microbiological characteristics in response to heavy metal contamination. Dehydrogenase
activity, soil ATP content, and enumeration of major soil microbial groups illustrated the effects of
contamination. There was a marked decrease in total culturable numbers of the different microbial groups.
In particular, asymbiotic nitrogen-fixers and heterotrophic bacteria were found to be sensitive.
Dehydrogenase activity was confirmed to be a good assay for determining the effect of heavy metals on
physiologically active soil microbial biomass.
The toxic effects of heavy metals on soil microorganisms are well known. However, less is known
about the relative sensitivity of different types of soil microorganisms (Rajapaksha et al., 2004).
Vaisvalavicius et al. (2006) assessed the toxicity of high concentrations of Pb (839 mg/kg), Zn (844
December 2008
9-117
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
mg/kg), and Cu (773 mg/kg) in the upper 0 to 0.1 m soil layer. Microbial abundance of all groups was
reduced and enzymatic activity was lower than for uncontaminated soil. In particular, actinomycetes,
oligonitrophobic and mineral N assimilating bacteria were most affected.
Effects of heavy metals in soil on microbes depends on soil pH, organic content, and the type of
heavy metal exposure (Kucharski and Wyszkowska, 2004). Some studies have shown that heavy metals
inhibit microbial activity in soil (Smejkalova et al., 2003; Vasundhara et al., 2004). However,
Wyszhowska et al. (2008) showed that heavy metals can either inhibit or stimulate the growth of soil
microbes. Populations of Azotobacter spp. decreased, but populations of oligotrophic and copiotrophic
bacteria, actinomyces, and fungi increased in response to heavy metal exposure. Acute metal stress causes
a decrease in microbial biomass as metal-sensitive microbes are inhibited (Joynt et al., 2006).
Soil OM cycling is known to be sensitive to disturbance due to heavy metal pollution. This can
cause increased litter accumulation at sites close to metal emissions point sources. The relative
importance of the various processes that might be responsible for this observation is poorly known.
Boucher et al. (2005) conducted C02 evolution studies in microcosms having metal-rich and metal-poor
plant materials. Their results suggested that there was a pool of less readily decomposable C that appeared
to be preferentially preserved in the presence of high metal (Zn, Pb, Cd) concentrations in the leaves of
the metallophyte Arabidopsis halleri. An additional possibility is that increased lignification of the cell
walls increased the amount of insoluble C (cf., Mayo et al., 1992).
Studies of the impacts of metal stress on the microbial community composition in soil have
generally been based on microbial culturing techniques that can select only a subset of the natural soil
population of microbes. More recent culture-independent studies have been conducted using
phospholipids or nucleic acid biomarkers to reveal information regarding changes in microbial
community structure (c.f., Joynt et al., 2006). Using this approach, Joynt et al. (2006) demonstrated that
soils contaminated with both metals (Pb, Cr) and organic solvent compounds over a period of several
decades had undergone changes in community composition, but still contained a phytogenetically diverse
group of bacteria. This may reflect adaptation to the potentially toxic conditions through such processes
as natural selection, gene exchange, and immigration.
Comparison between a severely contaminated soil with a similar soil that had much lower levels of
contamination showed considerably lower microbial diversity in the contaminated soil, particularly for
asymbiotic nitrogen fixers and heterotrophic bacteria (Oliveira and Pampulha, 2006).
Arbuscular mycorrhizal (AM) fungi can play important roles in mitigating toxicity of heavy metals
in plants. For example, AM symbiosis is known to be involved in plant adaptation to As-contaminated
soils. Higher plants that are adapted to As contaminated soils are generally associated with mycorrhizal
fungi (Gonzalez-Chavez et al., 2002). It has also been shown that AM symbioses can influence plant
December 2008
9-118
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
coexistence and community diversity (O'Connor et al., 2002). Some plants associated with AM fungi can
successfully colonize sites that are heavily contaminated by heavy metals (Pennisi, 2004).
Dong et al. (2008) cultivated white clover (Trifolium repens) and ryegrass (Lolium perenne) in
As-contaminated soil (water extractable As 82.7 mg/kg). The growth and P nutrition of both species
largely depended on AM symbiosis. The AM-inoculated plants showed selective uptake and transfer of P
over As.
9.6. Effects on Ecosystems
Because PM is heterogeneous with respect to chemical composition and size, it can cause a variety
of ecological effects, which were described by the U.S. EPA (2004) and by Grantz et al. (2003). These
effects are summarized below, based on those publications. Physical effects of particle deposition on
vegetation may include abrasion and radiative heating. Chemical effects may be more significant,
particularly from acidic particles associated with sulfate and nitrate (U.S. EPA, 2008e).
The effects of airborne particles are manifested via physical and chemical effects at the individual
organism (i.e., plant, microbe) level. However, individual organisms are interconnected within
populations, communities, and ecosystems. Ecosystems respond to stresses through their constituent
organisms. The responses of species and populations to atmospheric PM are determined by changes in
their physical and chemical environment that apply selection pressures on individual organisms. The most
common response in a vegetation community under stress is the elimination of the more sensitive
individuals and populations and an increase in abundance of those species that tolerate or are favored by
that particular stress.
Ecosystem response to pollutant deposition is a function of the ecosystem's ability to ameliorate
effects on individual plants and other organisms. At least three levels of biological interaction are
involved: (1) the individual organism and its environment, (2) the population and its environment, and (3)
the biological community composed of many species and its environment (Billings, 1978). Individual
organisms within a population vary in their ability to withstand the stress of environmental change. The
response of individual organisms within a population is based on their genetic constitution, stage of
growth at time of exposure to stress, and the microhabitat in which they are growing (Levin, 1998). The
range within which organisms can exist and function determines the ability of the population to survive.
Those able to cope with the stresses survive and reproduce. Competition among different species results
in succession (community change overtime) and, ultimately, produces ecosystems composed of
populations of species that have the capability to tolerate the stresses (Guderian, 1985; Rapport and
Whitford, 1999).
December 2008
9-119
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Ecosystems are subject to natural periodic stresses, such as drought, flooding, fire, and attacks by
biotic pathogens (e.g., fungi, insects). Natural perturbations return succession to an earlier stage; reduce
ecosystem structure and function; disrupt the plant processes of photosynthesis and nutrient uptake,
carbon allocation, and transformation that are directly related to energy flow and nutrient cycling; disrupt
food webs; and reduce the total nutrient inventory (Odum, 1993). Such transformations set the stage for
recovery and allow perturbed ecosystems to adapt to changing environments (Holling, 1986). Recovery
from natural perturbations can be rapid (Odum, 1993).
In contrast, anthropogenic stresses can result in damaged ecosystems that do not recover readily
(Odum, 1993; Rapport and Whitford, 1999). Ecosystems sometimes lack the capacity to adapt to
anthropogenic stresses and maintain their normal structure and functions unless the stressor is removed
(Rapport and Whitford, 1999). These stresses result in a process of ecosystem degradation marked by a
decrease in biodiversity, reduced primary and secondary production, and a lower capacity to recover and
return to its original state. In addition, there can be an increased prevalence of disease, reduced nutrient
cycling, increased dominance of exotic species, and increased dominance by smaller, short-lived
opportunistic species (Odum, 1985; Rapport and Whitford, 1999).
The possible effects of particulate (and other) air pollutants on ecosystems have been categorized
by Guderian (1977) as follows:
¦ accumulation of pollutants in plants and other ecosystem components (such as soil and
surface- and groundwater),
¦ damage to consumers as a result of pollutant accumulation,
¦ changes in species diversity because of shifts in competition,
¦ disruption of biogeochemical cycles,
¦ disruption of stability and reduction in the ability to self-regulate,
¦ breakdown of stands and associations, and
¦ expansion of denuded zones.
Ecosystem response to stress can be difficult to determine because the changes are often subtle.
This is particularly true of responses to atmospheric particles (U.S. EPA, 2004). Changes in the soil may
not be observed until pollutant deposition has occurred for many decades, except in the most severely
polluted areas around heavily industrialized point sources. The presence of co-occurring pollutants
generally makes it difficult to attribute ecological effects to PM alone or to one constituent in the
deposited PM. In other words, the potential for alteration of ecosystem function and structure exists but
December 2008
9-120
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
can be difficult to quantify except in cases of extreme levels of deposition, especially when there are other
pollutants present in the ambient air that may produce additive or synergistic responses.
9.6.1. Biogeochemical Processes
Atmospherically deposited PM can interact with a variety of biogeochemical processes.
Conclusions from the U.S. EPA (2004) are summarized here. In addition, there have been some more
recent modeling and mass balance studies that have attempted to quantify some of these linkages.
Atmospheric PM can affect ambient radiation, which can be considered in both its direct and
diffuse components. Foliar interception by canopy elements occurs for both up- and down-welling
radiation. Therefore, the effect of atmospheric PM on atmospheric turbidity influences canopy processes
both by radiation attenuation and by changing the efficiency of radiation interception in the canopy
through conversion of direct to diffuse radiation (Hoyt, 1978). Diffuse radiation is more uniformly
distributed throughout the canopy and increases canopy photosynthetic productivity by distributing
radiation to lower leaves. The enrichment in photosynthetically active radiation (PAR) present in diffuse
radiation appears to offset a portion of the effect of an increased atmospheric albedo due to atmospheric
particles.
The effects of regional haze on the yield of crops because of reduction in solar radiation were
examined by Chameides et al. (1999) in China, where regional haze is especially severe. They estimated
that approximately 70% of crops were being depressed by at least 3 to 5% by regional scale air pollution
and its associated haze (Chameides et al., 1999; U.S. EPA, 2004).
The ability of plants to take up metals from soil is an important part of metal cycling in the
environment. This uptake process allows the metals to enter the food web, where they might exert
mutagenic, carcinogenic, and teratogenic effects (Hunaiti et al., 2007). Some metals, including Cu, Co,
Ni, and Zn, are essential micronutrients needed for plant growth. Others, including Hg, Cd, and Pb are not
essential for plants. Kim et al. (2003) found decreased concentration of K in needles and Ca in stems of
Pinus sylvestris seedlings exposed to Cd addition. This response suggests a disturbance of nutrition in
response to Cd.
Pollutant-caused needle loss can reduce the interception of pollutants from the atmosphere, and
therefore reduce their concentrations in stemflow. This may be responsible for the observation that species
diversity of lichens is sometimes higher on trees affected by die-back (Hauck, 2003). This is an example
of pollution effects on trees actually reducing the extent of pollution effects on the lichens attached to
those trees.
December 2008
9-121
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
A number of mass balance approaches (Macleod et al., 2005; Toose and Mackay, 2004), and metal
speciation and transport models (c.f. Bhavsar et al., 2004a; 2004b; Gandhi et al., 2007) have been
developed in recent years.
9.6.2. Bioaccumulation
9.6.2.1. Metals
Biomagnification is the progressive accumulation of chemicals with increasing trophic level
(LeBlanc, 1995). Organic Hg is the most likely metal to biomagnify, in part because organisms can
efficient assimilate methylmercury and it is slowly eliminated (Croteau et al., 2005; Reinfelder et al.,
1998). In general, however, it has been assumed that metal biomagnification in aquatic ecosystems is an
exception rather than the rule (Gray, 2002). More recent research has demonstrated aquatic
biomagnification of certain metals. For example, (Stewart et al., 2004) used stable isotopes of C and N to
show biomagnification of Se in San Francisco Bay food webs. Croteau et al. (2005) identified trophic
position of estuarine organisms and food web structure in the delta of San Francisco Bay to document Cd
biomagnification in invertebrates that live on macrophytes and also in fish. Concentrations of Cd were
biomagnified 15 times within two trophic links in each food web. In contrast, no tendency towards
biomagnification was observed for Cu.
The study of trophic transfer and biomagnification is limited by the difficulty in discriminating
food webs and the uncertainty associated with assignment of trophic position to individual species
(Croteau et al., 2005). Use of stable isotopes can help to establish linkages. However, it is difficult to
determine the extent to which biomagnification occurs in a given ecosystem without thoroughly
investigating physiological biodynamics, habitat, food web structure, and trophic position of relevant
species. Thus, development of an understanding of ecosystem complexity is necessary to determine what
species might be at greatest risk from toxic metal exposure (Croteau et al., 2005).
Chlorinated POPs can be transported as particles through the atmosphere from industrial and
agricultural sources and deposited in remote regions. They have been detected in all levels of the Arctic
food web (Oehme et al., 1995). The polar bear is the top predator in the Arctic and feeds preferentially on
ringed seals and, to a lesser extent, on other seal species. Bioconcentration of organochlorines has been
shown in the Arctic food web, including fish, seals, and polar bears (Oehme et al., 1995).
Bioaccumulation of heavy metals can occur through the plant-herbivore and litter-detrivore food
webs. The U.S. EPA (2004) concluded that Cd and Zn can bioaccumulate in earthworms. Other
December 2008
9-122
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
invertebrates inhabiting soil litter may also accumulate metals. Although food web accumulation of a
metal may not result in mortality, it might reduce breeding potential.
Metal accumulation in litter can be found mainly around brass works and Pb and Zn smelters.
Organisms that feed on earthworms living in soils with elevated metal concentrations may also
accumulate Pb and Zn. Increased concentrations of heavy metals have been found in a variety of
mammals living in areas with elevated heavy metal concentrations in the soils.
Estuarine salt marshes are often located close to urban and industrial areas and receive elevated
levels of point and nonpoint (including atmospheric deposition) sources of trace metal contaminants.
Vegetation is important in the retention and accumulation of heavy metals in salt marshes. Soil
microorganisms, especially arbuscular mycorrhizal fungi (AMF), provide a key physical link between the
soil environment and plant roots. Carvalho et al. (2006) conducted experiments on the effects of AMF on
the uptake of Cd and Cu by Aster tripolium, a common plant species in polluted salt marshes and a host of
AMF. Carvalho et al. (2006) found that AMF colonization increased metal accumulation in the root
system of Aster tripolium without enhancing translocation to the shoot. By trapping toxic metals in the
roots, this plant species may reduce the extent of vegetative stress caused by metal exposure and act as an
effective sink for these metals.
Marine bivalve mollusks bioaccumulate trace metals and other contaminants (LaBrecque et al.,
2004) and therefore may be used as bioindicators of contamination. In addition, they constitute an
important link to human health by virtue of their importance as a food source (Cheggour et al., 2005; Li et
al., 2006).
There is not a standard method available for quantifying the bioavailability of heavy metals in soil.
A variety of models, isotopic studies, and sequential extraction methods have been used (c.f., Collins et
al., 2003; Feng et al., 2005; Shan et al., 2003). Total metal concentration in soil does not give a good
indication of potential biological effects because soils vary in their ability to bind metals in forms that are
not bioavailable. There are various methods available for assessing what is bioavailable, but soils are
heterogeneous and there is no ideal method for evaluating what conditions the soil biota experience.
Almas et al. (2004) argued that the actual measurement of biological effects is the best criterion for
determining bioavailability. In particular, the replacement of metal-sensitive microorganisms by
metal-tolerant organisms within each functional group may be one of the most sensitive indicators of
metal exposure. An increase in microbial trace metal tolerance per se would not be problematic if it was
not for the fact that this increase in tolerance is generally accompanied by a decrease in microbial
diversity (Almas et al., 2004; Lakzian et al., 2002).
Once metals accumulate to high concentrations in soils it is difficult to remove them. This is
because they are persistent and do not degrade. However, there are a variety of microbial and plant
December 2008
9-123
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
species that are known to accumulate high concentrations of metals when grown in metal-contaminated
soil (Prasad and De Oliveira Freitas, 2003). Plants that hyperaccumulate metals have potential for
remediation of metal-contaminated sites. About 400 species have been reported. Brassicaceae has the
largest numbers of taxa, with 11 genera and 87 species known to hyperaccumulate one or more metal
contaminants (Prasad and De Oliveira Freitas, 2003).
In aquatic ecosystems, biomagnification of trace metals does not necessarily occur. Nguyen et al.
(2005) found biodiminution for most metals in Lake Balaton, Hungary, with the exception of slight
enrichment of Zn from PM to zooplankton and of Cd from sediment to mussel.
Once transported to aquatic ecosystems, trace metals often preferentially bind to sediment particles.
Some of these sediment-bound metals may be unavailable to biota; in contrast, metals bound to sediment
organic matter may exhibit varying degrees of bioavailability (Di Toro et al., 2005). Piol et al. (2006)
studied the bioavailability of sediment-bound Cd to the freshwater oligochaete Lumbriculus variegatus.
They found that Cd uptake depended on the amount of free dissolved Cd(II), and the Cd contribution
from sedimentary particles to biological uptake was negligible.
9.6.2.2. Organics
The accumulation of PAHs in vegetation, due to their lipophilic nature, could contribute to human
and other animal exposure via food consumption. As a result, plant uptake of PAHs has been an important
area of research (Gao and Zhu, 2004). Most bioaccumulation of PAHs by plants occurs by leaf uptake
(Tao et al., 2006). Root uptake also occurs. It appears that roots preferentially accumulate the lower
molecular weight PAHs due to their greater water solubility (Wild and Jones, 1992).
Various models have been developed to simulate plant uptake of organic contaminants. The simple
partition-limited model of Chiou et al. (2001) has been further expanded to increase complexity and to
include root uptake pathways (e.g., Fryer and Collins, 2003; Yang and Zhu, 2007; Zhu et al., 2004).
In evaluating receptor choice for studies of contaminant exposure to plants, and also remediation
potential, it is important to consider differences among species. For example, Parrish et al. (2006)
assessed the bioavailability of PAHs in soil. During the first growing season, zucchini (Cucurbita pepo
ssp .pepo) accumulated significantly more PAHs than did other related plant species, including up to three
orders of magnitude greater levels of the six-ring PAHs. Parrish et al. (2006) also noted differences in
PAH uptake by two different species of earthworm.
Previously, there was relatively little information available regarding incorporation of
atmospherically deposited PAHs into aquatic food webs. It is known that PAHs can be transferred to
higher trophic levels, including fish, and that this transfer can be mediated by aquatic invertebrates, which
generally comprise an important part of fish diets. High mountain lakes offer an effective receptor for
December 2008
9-124
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
quantification of biomagnification in aquatic ecosystems from atmospheric PM deposition. There are
typically no sources of organic contaminants in their watersheds, and atmospheric inputs dominate as
sources of contamination. In addition, such lakes tend to have relatively simple food webs. Vives et al.
(2005) investigated PAH content of brown trout (Salmo trutta) and their food items. Total PAH
concentrations tended to be highest in organisms that occupy littoral habitats, and lowest in pelagic
organisms. This may reflect more efficient transfer of PAHs to underlying sediments in shallower water
and associated degradation within the water column.
It is difficult to discriminate between PAHs that are adsorbed to plant root surfaces as opposed to
those that are actually taken up by the roots. In general, soil bound PAHs are associated with soil OM and
are therefore not readily available for root uptake (Fismes et al., 2002; Jiao et al., 2007). Wild et al. (2005)
used two-photon excitation microscopy to visualize the uptake and transport of two PAHs (anthracene and
phenanthrene) from a contaminated soil into living wheat and maize roots. Jiao et al. (2007) developed a
sequential extraction method to discriminate between PAH adsorption in rice roots.
Maize roots and tops of plants have been shown to directly accumulate PAHs from aqueous
solution and from air in proportion to exposure levels. Root concentration factors are log-linear functions
of log-based octanol-water partition coefficients (log K0„); similarly, leaf concentration factors are
log-linear functions of log-based octanol-air partition coefficients (log Koa) (Lin et al., 2007). Although
the bulk concentrations of PAHs in various plant tissues can differ greatly, the observed differences
disappear after they are normalized to lipid content (Lin et al., 2007). This suggests that the lipid content
of different plant tissues may influence PAH distribution within the plant.
9.6.3. Nutrient Cycling
Upon entering the soil environment, PM pollutants can alter ecological processes of energy flow
and nutrient cycling, inhibit nutrient uptake, change ecosystem structure, and affect ecosystem
biodiversity. Many of the most important effects occur in the soil. The soil environment is one of the most
dynamic sites of biological interaction in nature. It is inhabited by microbial communities of bacteria,
fungi, and actinomycetes. These organisms are essential participants in the nutrient cycles that make
elements available for plant uptake. Changes in the soil environment that influence the role of the bacteria
and fungi in nutrient cycling determine plant and ultimately ecosystem response.
Many of the major indirect plant responses to PM deposition are chiefly soil-mediated and depend
on the chemical composition of the individual components of deposited PM. Effects may result in
changes in biota and in soil conditions that affect ecological processes, such as nutrient cycling and
uptake by plants.
December 2008
9-125
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
The soil environment is rich in biota. Bacteria and fungi are usually most abundant in the
rhizosphere, the soil around plant roots that all mineral nutrients must pass through. Bacteria and fungi
benefit from the nutrients that are present in root exudates and make mineral nutrients available for plant
uptake. The soil-mediated ecosystem impacts of PM are largely determined by effects on the growth of
bacteria and mycorrhizal fungi that are involved in nutrient cycling and plant nutrient uptake.
9.6.4. Ecosystem Structure and Function
Ecosystems are often subjected to multiple stressors, of which atmospheric PM deposition is only
one. Additional stressors are also important, including 03 exposure, climatic variation, natural and human
disturbance, the occurrence of invasive non-native plants, native and non-native insect pests, disease,
acidification, and eutrophication. PM deposition interacts with these other stressors to affect ecosystem
patterns and processes in ways that we are only beginning to understand.
Kiikkila (2003) investigated the effects of heavy metal pollution in proximity to a Cu-Ni smelter at
Harjavalta, Finland. The deposition of heavy metals increased within 30 km of the smelter. Only slight
changes in the understory vegetation were observed at distances greater than 8 km from the smelter. At 4
km distance, species composition of vegetation, insects, birds, and soil microbiota changed and tree
growth was reduced. Within about 1 km, only the most resistant organisms were surviving.
The Haijavalta region is one of the most intensively studied heavy metal polluted areas in the
world. Kiikkila et al. (2003) reviewed available data on heavy metal deposition and environmental effects
in this area. Emissions from the smelter were as high as 1100 t/yr of dust, 140 t/yr Cu, 96 t/yr Ni, 162 t/yr
Zn, and 94 t/yr Pb in 1987. Deposition levels decreased substantially after 1990, to only a few percent of
the amounts that occurred during the 1980s.
Tree growth (Scots pine) has been poor (Malkonen et al., 1999) and most vegetation was absent
within 0.5 km of the smelter. Effects on plant species occurrence close to the smelter were almost entirely
negative. In contrast, some animal species responded positively, including a leaf miner, three species of
aphid, and some ants, beetles, and spiders.
Inhibition of nutrient cycling and displacement by Cu and Ni of base cations from cation exchange
sites on the soil resulted in a decrease in base cation concentrations in the organic soil layer (Derome and
Lindroos, 1998; Kiikkila, 2003) close to the Harjavalta smelter. In addition, Mg, Ca, and Mn
concentrations in Scots pine needles were low, and this was attributed by Kiikkila (2003) to the toxic
effects of Cu and Ni to plant fine roots and also to ectomycorrhizal root tips (c.fi, Helmisaari et al., 1999).
Nutrient translocation during autumn was also affected close to the smelter; as a consequence needle
concentrations of K were relatively high (Nieminen et al., 1999).
December 2008
9-126
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
The number of soil animals clearly decreased and their community structure was altered close to
the Haijavalta smelter (Kiikkila, 2003). However, this effect was only pronounced within about 2 km of
the smelter. This suggests that the soil microfauna is relatively resistant to metal pollution effects.
Soil microbial activity decreased close to the Harjavalta smelter (Kiikkila, 2003), as reflected by
microbial respiration, distribution of species within physiological groups, and microbial and fungal
biomass. The fungi appeared to be more sensitive to metal contamination than the bacteria (c.f., Pennanen
et al., 1996). The rate of litter decomposition decreased, causing an accumulation of needle litter on top of
the forest floor near the smelter (c.f., Fritze et al., 1989).
Changes in breeding success of cavity-nesting passerine birds close to the Harjavalta smelter were
attributed by Kiikkila (2003) to habitat changes in response to metal toxicity. In particular, there was an
apparent decrease in the proportion of green insect larvae in the diet of nestlings. In addition, pollution
stress was inferred from increased heavy metals and decreased Ca in the diet of the pied fly catcher
(Ficedula hypoleuca) (Eeva et al., 2000).
As pollution levels increase, it is expected that the more sensitive species will be lost and the more
tolerant species remain. This gives rise to the concept of pollution-induced community tolerance (PICT),
which has been demonstrated for populations of bacteria and fungi (Davis et al., 2004). They assessed the
effects of long-term Zn exposure on the metabolic diversity and tolerance to Zn of a soil microbial
community across a gradient of Zn pollution. PICT was found to correlate better with total soil Zn than
with the concentration of Zn in soil pore water.
The toxicity of mixtures of metals is more difficult to determine than is the toxicity of a single
metal. The toxicity of the mixture might be approximately equal to the sum of the toxicities of individual
components. Alternatively, synergistic or antagonistic interactions can cause the toxicity of the mixture to
be lower or higher than the sum of the individual toxicities (Ince et al., 1999). There are a variety of
available approaches to assess the toxicity of mixtures of metals. For example, atoxic unit can be defined,
and the toxic units of individual contaminants summed, accounting for synergism and antagonism. The
toxic effect of a mixture can be described empirically as a polynomial function of individual toxic
concentrations, including cross terms. Alternatively, the attenuation of bioluminescence of a test
organism, such as Vibrio fischeri, can be measured (Utgikar et al., 2004).
9.7. Effects on Materials
Effects of air pollution on materials are related to both aesthetic appeal and physical damage.
Deposited particles, primarily carbonaceous compounds, cause soiling of building materials and culturally
important items, such as statues and works of art. Physical damage from dry deposition of PM also can
December 2008
9-127
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
accelerate natural weathering processes. The major deterioration phenomenon affecting building materials
in response to atmospheric deposition is probably sulphation, leading to secondary salt crystallization
which forms gypsum (Marinoni et al., 2003).
This section (a) summarizes information on exposure-related effects on materials associated with
particulate pollutants as addressed in the 2004 PM AQCD (U.S. EPA, 2004) and (b) presents relevant
information derived from very limited research conducted and published since completion of that
document. Most recent work on this topic has been conducted outside the U.S.
There are a variety of factors that contribute to the deterioration of monuments and buildings of
cultural significance. They include: (1) biodeterioration processes; (2) weathering of materials exposed to
the air; and (3) air pollution from both anthropogenic and natural sources (Herrera and Vide la, 2004).
Because of the diversity in climate, proximity to marine aerosol sources, and pollution of various types,
the magnitude and relative importance of these causal agents vary by location.
Much existing literature on damage to structural materials of cultural heritage has not seriously
considered the importance of biodeterioration process and the relationship that often exists between
environmental characteristics and the microbial communities that colonize monuments and buildings. In
general, high humidity, high temperature, and air pollution often enhance the biodeterioration hazard.
Herrera and Videla (2004) concluded that heterotrophic bacteria, fungi, and cyanobacteria were the main
microbial colonizers of buildings that they investigated in Latin America. Their analyses suggested that
the major deterioration mechanism of limestone at the Mayan site of Uxmal in a non-polluted rural
environment was biosolubilization induced by metabolic acids produced by bacteria and fungi. The rock
decay at Tulum, near the seashore, was mainly attributed to the marine influence. At Medellin, it appeared
that biodeterioration effects from microbes synergistically enhanced the effects of atmospheric factors on
material decay. Deterioration of structural material in the Cathedral of La Plata, located in a mixed
urban/industrial environment, was attributed mainly to atmospheric pollutants (Herrera and Videla, 2004).
Ambient particles can cause soiling of man-made surfaces. Soiling generally is considered an
optical effect. Soiling changes the reflectance from opaque materials and reduces the transmission of light
through transparent materials. Soiling can represent a significant detrimental effect, requiring increased
frequency of cleaning of glass windows and concrete structures, washing and repainting of structures,
and, in some cases, reducing the useful life of the object. Particles, especially carbon, may also help
catalyze chemical reactions that result in the deterioration of materials during exposure (U.S. EPA, 2004).
Soiling is dependent on atmospheric particle concentration, particle size distribution, deposition
rate, and the horizontal or vertical orientation and texture of the exposed surface (Haynie, 1986). The
chemical composition and morphology of the particles and the optical properties of the surface being
December 2008
9-128
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
soiled will determine the time at which soiling is perceived by human observers (Nazaroff and Cass,
1991).
Ferm et al. (2006) reported development of a simple passive particle collector for estimating dry
deposition to objects of cultural heritage. The observed mass of deposited particles mainly belonged to the
coarse particulate mode. The sampler collects particles from all directions. It replicates at least some of
the complexity of particle deposition to actual objects, and is easier to analyze than a precious object
(Ferm et al., 2006).
Soiling of urban buildings constitutes a visual nuisance that leads to the loss of architectural value.
Soiling can include reversible darkening of the building surfaces and also irreversible damage. Water
runoff patterns on the building surfaces are influenced by the type of surface material, architectural
elements, and climate. Therefore, soiling does not occur uniformly across the building. Public perception
of soiling entails complex interactions between the extent of soiling, architecture, and aesthetics (Grossi
and Brimblecombe, 2004).
One of the most significant air pollution damage features affecting urban buildings and monuments
is the formation of black crusts. Quantification of different forms of carbon in black crusts is difficult.
There is often a carbonate component which is derived from the building material, plus organic carbon
and EC, derived from air pollution. Elemental C is considered to be a tracer for combustion sources,
whereas organic C may derive from multiple sources, including atmospheric deposition of primary and
secondary pollutants, and the decay of protective organic treatments (Bonazza et al., 2005). Bonazza et al.
(2005) quantified organic and elemental C in damage layers on European cultural heritage structures.
Organic C predominated over elemental C at almost all locations investigated. Traffic appeared to be the
major source of fine carbonaceous particles, with organic matter as the main component (Putaud et al.,
2004). Viles and Gorbushina (2003) found that soiling in Oxford, UK showed a relationship with traffic
and N02 concentrations.
In addition to the soiling effects of BC, much soiling appears to be largely of microbiological origin
(Viles and Gorbushina, 2003). Microbial biofilms, composed mainly of fungi, can stain exposed rock
surfaces with yellow, orange, brown, gray, or black colors. Microorganisms may be able to trap PM more
efficiently than the stone surface itself. In addition, microbial growth may be stimulated by organic or
nutrient constituents in PM deposition.
Viles et al. (2002) investigated the nature of soiling on limestone tablets in relation to ambient air
pollution and climate at three contrasting sites in Great Britain over periods of one to eight years.
Spectrophotometer and microscope observations suggested that there were not consistent trends in soiling
over time at the study sites. Each site behaved differently in terms of the temporal development of soiling
December 2008
9-129
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
and the differences between sheltered and exposed limestone tablets. In addition, organisms played
important roles in the soiling response, even at the highly polluted site.
Some work has been conducted on public perception regarding the lightness of historic buildings
and the aesthetic need for cleaning subsequent to soiling by air pollution. Brimblecombe and Grossi
(2005) found a strong relationship between the perceived lightness of a building and the opinion that it
was dirty. This relationship was used to establish levels of blackening that might be publicly acceptable.
Recently, the importance of organic contaminant deposition to the overall air pollution damage to
building materials has been recognized. Low molecular weight organic anions such as formate, acetate,
and oxalate are ubiquitous in black crusts in damage layers on stones and mortars sampled from
monuments and buildings throughout Europe (Sabbioni et al., 2003). This has been observed at urban,
suburban, and rural sites.
9.7.1. Effects on Paint
Studies have evaluated the soiling effects of particles on painted surfaces (U.S. EPA, 2004).
Particles composed of EC, acids, and various other constituents are responsible for the soiling of
structural painted surfaces. Coarse-mode particles (>2.5 |im) initially contribute more soiling of
horizontal and vertical painted surfaces than do fine-mode particles (<2.5 ^m), but are more easily
removed by rain (Haynie and Lemmons, 1990). Rain interacts with coarse particles, dissolving the
particle and leaving stains on the painted surface (Creighton et al., 1990; Haynie and Lemmons, 1990).
Particle deposition contributes to increased frequency of cleaning of painted surfaces and physical
damage to the painted surface. Air pollution affects the durability of paint finishes by promoting
discoloration, chalking, loss of gloss, erosion, blistering, and peeling (U.S. EPA, 2004). There have been
no new developments in this field subsequent to the review of the U.S. EPA (2004).
9.7.2. Effects on Metal Surfaces
Metals undergo natural weathering processes. The effects of air pollutants on natural weathering
processes depend on the nature of the pollutant(s), the deposition rate, and the presence of moisture
(U.S. EPA, 2004). Pollutant effects on metal surfaces are governed by such factors as the presence of
protective corrosion films and surface electrolytes, the orientation of the metal surface, and surface
moisture. Surface moisture facilitates particulate deposition and promotes corrosive reactions. Formation
of hygroscopic salts increases the duration of surface wetness and enhances corrosion.
December 2008
9-130
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
A corrosion film, such as for example the rust layer on a metal surface, provides some protection
against further corrosion. Its effectiveness in retarding the corrosion process is affected by the solubility
of the corrosion layer and the pollutant exposure. Other than the effects of acidifying compounds, there
has not been additional research conducted in recent years on the effects of PM deposition on metal
corrosion.
9.7.3. Effects on Stone
Air pollutants can enhance the natural weathering processes on building stone. The development of
crusts on stone monuments has been attributed to the interaction of the stone's surface with pollutants,
wet or dry deposition of atmospheric particles, and dry deposition of gypsum particles. Because of a
greater porosity and specific surface, mortars have a high potential for reacting with environmental
pollutants (Zappia et al., 1998).
Most research evaluating the effects of air pollutants on stone structures has concentrated on
gaseous pollutants (U.S. EPA, 2004). The dark color of gypsum is attributed to soiling by carbonaceous
particles. A lighter gray colored crust is attributed to soil dust and metal deposits (Ausset et al., 1998;
Camuffo, 1995; Lorusso et al., 1997; Moropoulou et al., 1998). Lorusso et al. (1997) attributed the need
for frequent cleaning and restoration of historic monuments in Rome to exposure to total suspended
particulates.
Grossi et al. (2003) investigated the black soiling rates of building granite, marble, and limestone in
two urban environments with different climates. Horizontal specimens were exposed, both sheltered and
unsheltered from rainfall. Limestone showed soiling proportional to the square root of the time of
exposure, but granite and marble did not.
Black soiling is caused mainly by particulate EC (PEC). For that reason, it is most prevalent in
urban environments due to the formation of carbonaceous fine particles from the incomplete combustion
of fossil fuels. Traffic emissions, especially from diesel engines, and wood burning are important sources
of PEC (Grossi et al., 2003).
Kamh et al. (2005) studied the effects of weathering on Conway Castle, an historical structure in
Great Britain built about 1289 AC. The weathering was identified as honeycomb, blackcrust, exfoliation,
and discoloration, with white salt efflorescence at some parts. These features are diagnostic for salt
weathering (Goudie et al., 2002), and this was confirmed by laboratory analyses, including scanning
electron microscopy and x-ray diffraction. The authors concluded that the salt was derived from three
sources: sea spray, chemical alteration of the carbonate in mortar into sulfate salts by acidic deposition,
and wet deposition of air pollutants on the stone surface. The salt content on the rock surface fills the rock
December 2008
9-131
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
pores and then exerts high pressure on the rock texture due to hydration of the salt in the cold humid
environment. In particular, CaS04 and Na2S04 exert enough pressure on hydration as to deteriorate
construction rock at both the micro- and macroscale (Moses and Smith, 1994).
9.8. Effects on Climate
The Intergovernmental Panel on Climate Change (IPCC) summary documents Climate Change
2007 provided substantial background information on the effects of PM on climate. Much of the material
presented in this section is taken from these documents, especially from Chapter 2 (Changes in
Atmospheric Constituents and in Radiative Forcing) of the Working Group I Report, The Physical Science
Basis (IPCC, 2007). The reader is referred to these reports for more detailed discussion of the topics
summarized here.
Ecosystems across the Earth are tightly interconnected. Changes in one compartment can affect the
state of another. Feedbacks between compartments can amplify or mitigate changes (Grannas et al.,
2007). Such connections apply to the climate system. PM in the atmosphere adsorbs and scatters solar
radiation, interacts with water, and changes cloud properties. Therefore, atmospheric PM can affect
multiple aspects of climate (Seinfeld and Pankow, 2003). Fine-mode aerosols have sizes close to the
wavelengths of visible light, and therefore are expected to have a stronger effect on climate than larger
particles. Also, fine-mode particles can be transported far from their source region and therefore
contribute to spatially diffuse effects (Kanakidou et al., 2005).
The radiative forcing of climate by most anthropogenic aerosols has long been known to be
opposite in sign to effects from greenhouse gases(GHG) (Mahlman, 1997). The major aerosol
components include sulfates and other inorganic species and carbonaceous species (Park, 2005 #4735).
The carbonaceous species include complex mixtures of organic compounds plus BC or soot though the
composition of the mix of organic compounds is poorly characterized.
PM has long been known to have important effects that directly and indirectly modifying the
climate system (Aunan et al., 2006; Hansen and Sato, 2001; Heywood and Shine, 1995). Atmospheric
particles like S042 and organic carbon cool the atmosphere through scattering of shortwave radiation;
others like BC warm the atmosphere through absorption of shortwave radiation. In addition, PM can
influence climate by affecting clouds and the albedo of snow and ice (Aunan et al., 2006). Such effects
appear to be especially pronounced in the developing world. For example, Menon et al. (2004) suggested
that the observed trend towards increased summer floods in southern China and drought in northern China
might be linked to radiation-absorbing atmospheric particles.
December 2008
9-132
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
The review of Poschl (2005) provided an overview of the current state of knowledge, major
uncertainties, and research perspectives on the properties and interactions of atmospheric aerosols and
their effects on climate. Airborne PM undergoes a variety of physical and chemical interactions and
transformations that result in changes in particle size, structure, and composition (Figure 9-54). Clouds
are formed by condensation of water vapor on pre-existing aerosol particles (cloud condensation nuclei
[CCN] and ice nuclei [IN]). Most clouds re-evaporate, releasing the aerosol particles back into the
atmosphere. If the cloud particles form precipitation, the aerosol particles, including CCN, are scavenged
from the atmosphere. Depending on aerosol properties and meteorological conditions, the lifetime of
aerosol PM in the atmosphere ranges from hours to weeks (Poschl, 2005). There is substantial uncertainty
in the magnitude of the aerosol effects, and much of this uncertainty concerns changes in cloud properties.
Figure 9-54. Atmospheric cycling of aerosols.
There is a complex interdependence of composition, composition-dependent properties,
atmospheric interactions and transformations, climate effects, and aerosol sources (Figure 9-55) (Figure
E&S-7, Poschl, 2005). The associated feedback loops constitute areas of intense research central to
climate science.
SaooncteKy
! nutmmn
I
4 / CloudProosss.mf} \\
Pttpfeaf.a«iCftemleal Aging i'a
Nature
Anthropogenic
Wet
Primer*
>'t I >GH
Source: Poschl (2005)
December 2008
9-133
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Aarosol Composition
* .''{jrTiCXsr ma rriaiS CxwiL-fJnci ofrun, bllu
(toWWWt Sl*f#jt», SfrUCfWB, M( crf»«tfp»
¦ wii-pasiTicm iVwrtcte mxfgmptmet
Aerosol Source®
• pnmuy wmastem mi secondary fomaum
»rahaat* antf arti'fwopegeriit
I -»rt|u»Wy cerifrti1 antf mgtmom
¦ cn»ifea> reactiwiy mit»otogicai ac#wl*
¦ ft.rsi'oa^eiofj', CCN #ncf 94 m&oty
; Interactions & TrwwfOrnMticii
* M¥$
# wv
¦ cfewls Bftci pmTifiitglkm
• mmir budget, fiytboiogtcal cycto, um0mm
¦ reS&'MOfV. Mf&Ctt&fS.,
mwi
Rfuwj, Interdependence and feedback between atmospheric, aerosol composition, properties, interactions
and transformation, climate and health effects, and sources.
Source: Poschl (2005).
Figure 9-55. Interdependence and feedback between atmospheric aerosol composition, properties,
interactions and transformation, climate and health effects, and sources.
The IPCC Third Assessment Report (IPCC, 2001) categorized radiative forcings (RFs) from
aerosols into direct and indirect effects (see Figure 9-56). The direct effect involves scattering and
absorption of shortwave and longwave radiation, directly altering the radiative balance of the
Earth-atmosphere system. Multiple aerosol components contribute to this direct effect, including sulfate,
fossil fuel organic C, fossil fuel black C, biomass burning, and mineral dusts (Menon et al. 2002; Wang
2007). Scattering aerosols exert a net negative (cooling) RF. Partially absorbing aerosols can exert a net
negative top-of-the-atmosphere (TOA) RF over dark surfaces like oceans and dark forests, or a net
positive TOA RF over bright surfaces, such as desert, snow, ice, or cloud fields (Chylek and Wong, 1995;
Forster et al., 2007; Haywood and Shine, 1997).
The indirect effects involve modification of the radiative properties, amount, and lifetime of clouds.
They are determined by the effectiveness of aerosol particles to act as CCN (Forster et al., 2007; Penner,
2001). The component of the indirect effect that pertains to the cloud droplet number concentration and
cloud droplet size is called the "first indirect effect" or "cloud albedo effect." The component that pertains
to cloud liquid water content, cloud height, and lifetime is called the "second indirect effect" or the "cloud
lifetime effect" (Forster et al., 2007; Lohman and Feichter, 2005; Ramaswamy et al., 2001).
December 2008
9-134
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Natural and Afithropogcn c Cruras oris cf Pjrt'r.ci and Precursor Gases
* sort' tuirwraf and road dust; St j s> i >, »>.« ,»'A »„» .. . >jO ««
i
Indirect Etfmts
Direct
Effects
COV and IN mmly
otero. and MoT. aanniy
a&M»|*fcw and scattering (mf,
solar (UVAflS) and ionwsfmi
I ~__L
Atmospheric and Oceanic Circulation, lliogaoetiemical Cycles
¦ mam mitt tmal transport, horizontal arid verticBl trrnnport mtmim weather enente
¦ meim cycle' surface *nd grown? mter, mmantflet.
* carton ma sulfur cycle* ptmimynShesis and decay a* tifflnss, fiiotogfcal rmtaboSs.
biomass burring and Kmi fuel eomhiisiim; votmvm, ate.
*
Source: Poschl (2005).
Figure 9-56. Direct and indirect aerosol effects and major feedback loops in the climate system.
Figure 9-56 illustrates these effects and some of the major feedback loops. The key feedback loops
involve various interactions of atmospheric aerosols with solar and terrestrial radiation, clouds and
precipitation, general circulation, hydrological cycle, and with natural and anthropogenic aerosol sources.
Each of the interactions shown in Figure 9-56 comprises a wide range of poorly understood processes
which are not well quantified. Moreover, the actual climate system responses are highly variable and
uncertain. In many cases, even the sign or direction of the effect is not known. For example, increased
C02 is expected to increase photosynthesis, biogenic emissions of VOC, and the formation of SOA
particles which can serve as CCN, increase cloudiness, and cause cooling (Kanakidou et al., 2005). On
the other hand, this effect could be countered by temperature-related biological stress, eutrophication,
decreased photosynthesis, decreased VOC emissions and secondary organic aerosol formation and
cloudiness, and enhanced warming (Poschl, 2005). Such complexities are among the most important
reasons for the uncertainty of climate sensitivity and for the moderately wide range of projected global
mean surface temperature increases over the next century (Kanakidou et al., 2005; Lohman and Feichter,
2005; Poschl, 2005).
December 2008 9-135 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
9.8.1. Direct Effects
9.8.1.1. Radiation Budget
RF is quantified as the rate of energy change per unit area of the earth as measured at the top of the
atmosphere in units of watts per square meter (W/m2). This allows quantitative comparison of various
natural and anthropogenic climate drivers. The total direct aerosol RF estimated by models and
observations is on the order of -0.5 (± 0.4) W/m2, with a medium-low level of scientific understanding
(Forster et al., 2007). The negative sign indicates an overall direct cooling effect from aerosols. The direct
RF of individual aerosol species including sulfate, fossil fuel OC, BC, biomass burning, nitrate, and
mineral dust are less certain than the estimated total direct aerosol RF. The estimated direct RF of -0.5
W/m2 for total aerosols compares with an RF of+1.66 W/m2 for C02 and +0.48 W/m2 for CH4 (Forster et
al., 2007).
Atmospheric PM can alter the characteristics and net receipt of solar radiation. The characteristics
and amounts of environmental radiation in turn are important in determining rates of photosynthesis and
water cycling. Atmospheric turbidity describes the degree of scattering occurring in the atmosphere due to
particles and gases. Total particle-based extinction, however, is the sum of both scattering and absorption.
Absorption of short-wavelength solar radiation reduces the amount of radiation reaching the Earth's
surface and leads to atmospheric heating (U.S. EPA, 2004). If the absorbing particles re-radiate in the
infrared range, some of this energy is lost as long-wave re-radiation to space. The balance of this energy is
captured at the surface as down-welling infrared radiation. Canopy temperature and transpirational water
use by vegetation are particularly sensitive to long-wave, infrared radiation. Atmospheric heating caused
by particles in the atmosphere reduces vertical temperature gradients, and this could reduce the intensity
of atmospheric turbulent mixing. The magnitude of such potential effects on turbulent transport is
unknown.
Atmospheric turbidity increases the intensity of diffuse (sky) radiation. In a clear atmosphere,
diffuse radiation may be on the order of 10% of total solar radiation, whereas under highly turbid humid
conditions this fraction will be higher. The ratio is highest at solar noon and lowest near dawn or dusk
when the path length through the atmosphere is longest. The wavelength dependence of particle scattering
induces an enrichment of photosynthetically active radiation (PAR) with respect to total or direct beam
radiation (U.S. EPA, 2004).
Aerosols produced by incomplete combustion, including those produced by forest fire and fuel
combustion, contain significant fractions of BC which absorbs across the solar and terrestrial radiation
spectra. The presence of absorbing aerosols reduces the ratio of PAR to total radiation received at the
surface, potentially reducing photosynthetic water uptake efficiency. The net effect of aerosol absorption
December 2008
9-136
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
on the surface depends on the relative magnitudes of the particulate absorption coefficients in the visible
and infrared area and on the albedo of the Earth's surface (U.S. EPA, 2004).
The largest effects of atmospheric PM on visibility and atmospheric turbidity are due to light
scattering. Non-absorbing, scattering aerosols raise the overall albedo of the atmosphere and reduce the
amount of radiation reaching the Earth's surface by the amount reflected or scattered back into space.
U.S. EPA (2004) reported the results of analysis of data collected by a global network of thermopile
pyranometers operated by the World Meteorological Organization indicating a 50-year global reduction of
2.7% per decade in the amount of solar radiation reaching the Earth's surface. This has been associated
with an increasing global albedo caused by an increasing abundance of atmospheric particles.
The absorption of solar radiation by atmospheric particles, together with the trapping of infrared
radiation emitted by the Earth's surface by certain gases, enhances the heating of the Earth's surface and
lower atmosphere. It is generally believed that greenhouse gas emissions caused most of the global mean
warming observed during the 20th century and that S042 and other aerosols counteracted this warming to
some extent by reflecting solar radiation to space and therefore cooling the lower atmosphere and the
Earth's surface. The role of BC in absorbing incoming shortwave radiation and thereby heating the lower
atmosphere may have been underestimated (Hansen and Sato, 2001; Novakov et al., 2003). However, a
detection and attribution analysis by Jones et al. (2005), with recent Hadley Centre simulations (Roberts
and Jones, 2004), examined the sensitivity of the effects of greenhouse gas variations on climate with the
inclusion of BC. They found that the S042 pattern was too similar to the inverse of the BC pattern to be
able to deduce the contribution to warming over the past 100 years due to BC. There was no evidence that
BC counteracted to any substantial degree the cooling effect of S042 aerosol.
The principal aerosols that can influence climate are listed in Table 9-4, along with their primary
sources and the estimated ranges of contribution computed using different models (Haywood and
Boucher, 2000; IPCC, 2001). OC aerosols are mostly reflective. BC aerosols are mostly absorbing in the
visible and UV regions, and this causes BC to counteract the cooling caused by other aerosols. The
amount of BC emitted by a combustion source depends on the burning efficiency of the fuel as well as the
mass of fuel burned. The mass of BC emitted per unit mass of fuel burned (emission factor) depends on
fuel type (i.e., coal, biomass, diesel) (i.e., coal, biomass, diesel; Menon, 2004). For example, coal burning
in an inefficient stove or furnace produces orders of magnitude more BC than the same fuel burned in a
modern electric power plant (Menon, 2004).
Smoke particles that originate from fires affect climate primarily by scattering and absorbing solar
radiation. This direct radiative forcing reduces the net solar radiation that reaches the Earth's surface
(Davison et al., 2004; Kobayashi et al., 2004). In addition, soot aerosols increase air temperature, reduce
relative humidity, and reduce cloud cover (Liu, 2005). McConnell et al. (2007) investigated the deposition
December 2008
9-137
DRAFT-DO NOT QUOTE OR CITE
-------
1 history of BC in Greenland ice cores and found that deposition peaked during the industrial revolution,
2 resulting in climate forcing of 3 W/m2 from 1906 to 1910. The estimated median surface radiative forcing
3 in early summer based on the ice coring was 0.42 W/m2 before 1850, 1.13W/m2 during the period from
4 1850 to 1951, and 0.59 W/m2 after 1951.
Table 9-4. Range in estimated source strength (Tg aerosol/year).
Type
Source Strength
SULFATE
Industrial
65-92.4
Ocean
10.7-23.7
Aircraft
0.04
Biomass burning
2.03.0
ORGANIC CARBON
Fossil fuel
10-20
Biomass
30-45
BLACK CARBON
Fossil fuel
5.8-6.6
Biomass
6.0-17.2
NITRATES
Fossil fuel
0.3
Biomass
5.7
Other (human, soils, animal, agriculture)
74.5
SEA SALT
< 2 pm 82
>2 pm
2583
DUST
< 2 pm
243
E
C\J
A
Source: Menon (2004)
5 Much interest has developed in defining more precisely the role of pyrogenic C in the boreal C
6 cycle. This is due to: (1) the resistance of pyrogenic C to decomposition; (2) its influence on soil
7 processes; and (3) the absorption of solar radiation by soot aerosols (Preston and Schmidt, 2006).
8 Preston and Schmidt (2006) reviewed the current state of knowledge regarding atmospheric emissions of
9 pyrogenic C in the boreal zone. They considered chemical structures, analytical methods, formation,
10 characteristics in soil, loss mechanisms, and longevity. Biomass is largely converted to gaseous forms
December 2008
9-138
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
during burning, but up to several percent is converted to pyrogenic C, and this includes charcoal and BC.
Charcoal is defined visually; BC is defined chemically by its resistance to oxidation in the laboratory.
Andreae and Gelencser (2006) reviewed a different category of light-absorbing carbon, referred to
as brown carbon. Operational methods used to measure all light-absorbing carbon are estimated to have a
factor-of-2 at present.
Within the boreal zone, fire is a critical driver of ecosystem process and nutrient cycling (Hicke et
al., 2003). For example, Bachelet et al. (2005) estimated that 61% of the C gained in Alaska by primary
production of boreal forests between 1922 and 1996 was lost to fire.
An updated modeling effort to evaluate the radiative effects of aerosols was presented by Stier et al.
(2007). Inclusion of refractive indices recommended by Bond and Bergstrom (2005) significantly
increased aerosol RF and resulted in better agreement with sun-photometer estimates. Although this stage
of climate modeling improved the representation of aerosols, large uncertainties remain regarding the
effects of aerosol mixing and aerosol-cloud interactions. Furthermore, Stier et al. (2007) emphasized that
these types of modeling efforts are dependent upon emission estimates that are likely to vary by a factor
of 2 or more.
One important reason for the acknowledged uncertainty in estimating global emissions of
carbonaceous aerosols is the influence of intermittent fires that can occur at scales large enough to affect
hemispheric aerosol concentrations. To better quantify the effects of large-scale fire, Generoso et al.
(2007) used satellite observations of boreal fires in Russia in 2003 to evaluate the performance of a global
chemistry and transport model in simulating aerosol optical thickness, transport, and deposition.
Emissions estimates of BC and OC were adjusted in the model to better match satellite observations of
pollutant transport over the North Pacific. This resulted in an increase in optical thickness and BC
deposition by a factor of 2. The adjusted model estimated that the fires contributed 16-33% of the optical
thickness and 40-56% of BC deposition north of 75° N in the spring and summer of 2003.
Large fires also occurred over the Iberian Peninsula and Mediterranean coast during 2003. A
meso-scale atmospheric transport model was used with ground-based measurements and satellite optical
measurements to characterize the dispersion of emitted smoke particles and quantify radiative effects
across Europe (Hodzic et al., 2007). The modeled wildfire emissions resulted in increases in PM2.5
concentrations from 20 to 200%. The increased aerosol concentration was estimated to increase radiative
forcing by 10-35 W/m2 during the period of strong fire influence. Absorption of radiation by BC was also
estimated to decrease rates of photolysis by 30%. In this simulation, all particles were assumed to be
internally mixed, and secondary aerosol formation was not considered. Meteorological conditions in
Europe during the exceptionally hot summer of 2003 were linked to enhanced photochemically derived
December 2008
9-139
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
pollutants, increased wild fires, and elevated aerosol concentrations in an analysis by Vautard et al.
(2007a).
In addition to incidental fires, routine biomass burning, usually associated with agriculture in
eastern Europe, also has been shown to contribute to hemispheric concentrations of carbonaceous
aerosols. In the spring of 2006, the most severe air pollution levels in the Arctic to date were recorded
(Stohl et al., 2007). Atmospheric transport modeling coupled with satellite fire detection data identified
biomass burning for agriculture as the primary cause of the high pollution levels. Concentrations of PM25
peaked during the pollution episode at values of an order of magnitude greater than those recorded prior
to the episode. The increased transport of pollution into the Arctic during 2006 was attributed to weather
conditions that delayed preparations for crop planting into May. Weather patterns favorable for pollutant
transport into the Arctic were related to unusually warm weather in late April and May, when the majority
of agricultural biomass burning took place that year.
In the summer of 2004, 2.7 million ha were burned by wildfire in Alaska and 3.1 million ha were
burned in Canada. Effects on atmospheric air quality were measured throughout the Arctic, although the
concentrations of particulates varied considerably. Aerosol optical depths were also increased at all
measurement stations, which indicated that the fires were likely to have had a significant effect on the
atmospheric radiation budget for the Arctic (Stohl et al., 2006). At one site, a pronounced drop in albedo
was observed due presumably to high deposition of light absorbing particulates on the snow surface by
the North American fires in 2004.
Investigations of the effects of large fires on climate forcing have typically focused on the
absorptive effects of BC. However, these fires also release large amounts of C02 and CH4, as well as light
scattering compounds such as OC, and can enhance cloud formation. These fires also increase radiative
surface absorption through BC deposition on snow and ice, and alter surface albedo and ecosystem energy
budgets within the burn perimeter. Randerson et al. (2006) estimated the net climate forcing of
greenhouse gases, aerosols, BC deposition on snow and ice and changes in albedo for the year subsequent
to a fire and for 80 years in the future in interior Alaska. The net effect of the fire in the first year was an
increase in radiative forcing, but over the 80-year recovery period, average net annual radiative forcing
was decreased by the fire.
International shipping has been identified as an additional source of carbonaceous aerosols.
Simulations with a climate model that included aerosol effects and 3 different emissions inventories
showed that shipping contributed 2.3-3.6% of the total sulfate atmospheric aerosol content and 0.4-1.4%
of the total BC atmospheric aerosol content, based on global means in 2000. This modeling also showed
that aerosol optical thickness over the Indian Ocean, the Gulf of Mexico, and the northeastern Pacific
Ocean varied by 8 to 10%. The corresponding all-sky (that includes both cloudy and clear skies) direct
December 2008
9-140
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
radiative forcings ranged from -0.011 W/m2 to -0.013 W/m2. The greatest effect of aerosols emitted from
global shipping is likely to be an increase in cloud formation and the resulting change in reflectivity of
shortwave radiation. Aerosols from shipping were estimated to contribute 17-39% of the total
anthropogenic aerosol radiation forcing effect.
9.8.1.2. Temperature
BC is the main component of soot in the atmosphere and it changes the temperature of the air in
three ways (Jacobson, 2004): Daytime warming of the atmosphere and cooling of the ground surface
immediately below the BC PM. This atmospheric warming is caused by absorption of solar radiation by
the soot particles. Because this radiation does not reach the ground, the ground is cooled; Nighttime
warming of both the atmosphere and the ground due to absorption of the Earth's thermal-infrared
radiation, a portion of which is redirected back to the ground. This warming also occurs primarily locally,
in the vicinity of the soot particles; Large-scale daytime and nighttime warming of the air by advected
molecules in the atmosphere that had previously been heated by the soot. These warmed molecules can
have long lifetimes and be transported large distances remote from the soot PM that initiated the
warming.
When BC is deposited to the surface of ice or snow, solar absorption and heating occur at the
surface. This can melt additional snow or ice at the surface and the reflectivity of the surface can change.
Both factors affect aspects of climate. Jacobson (Jacobson, 2003a, 2004) and Jacobsen et al. (Jacobson,
2003c) estimated the warming due to fossil fuel BC and organic matter using the Gas, Aerosol, Transport,
Radiation, General Circulation, Mesoscale and Ocean Model (GATOR-GCMOM). The modeling effort
included consideration of the BC cycle, accounting for emissions, transport, aerosol coagulation, aerosol
growth, cloud activation, aerosol-cloud coagulation, cloud-cloud coagulation, rainout, washout, dry
deposition, and processes of precipitated and dry-deposited BC in snow and sea ice. Results suggested
that BC absorption in snow and sea ice increased near-surface temperatures over a 10-year simulation by
about 0.06°K (Jacobson, 2003c).
BC soot is a potentially important agent of climate warming in the Arctic, and northern boreal
wildfires may contribute substantially to this effect. Soot is approximately twice as effective as C02 in
altering surface air temperature, and can reduce sea ice formation and snow albedo (Hansen and
Nazarenko, 2004).
Kim et al. (2005c) investigated the relationships between northern boreal wildfires and reductions
in Arctic sea ice and glacial coverage. They modeled the FROSTFIRE boreal forest control burn
(Hinzman et al., 2003) with respect to BC aerosol transport, dispersion, and deposition. Model results
suggested that boreal wildfires could be a major source of BC soot to sea ice and glaciers in Alaska. This
December 2008
9-141
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
may exacerbate summer melting of sea ice and reduce recruitment of first year ice into multi-year ice,
thereby leading to an overall reduction in sea ice. Similarly, increased BC soot on glaciers would be
expected to increase summer melting and lead to an overall reduction in glacial coverage (Kim et al.,
2005c).
Jacobson (2002b) proposed, based on model simulations with 12 identifiable effects of aerosol
particles on climate, emission reductions of fossil fuel particulate BC and associated organic matter could
potentially slow warming for a specific period more than reduction of C02 or CH4 for a specific period.
Jacobson's (2006) calculations suggested that fossil fuel BC plus organic matter emissions reductions
could eliminate 8 to 18% of total anthropogenic warming, and 20 to 45% of net warming after accounting
for aerosol cooling, within a period of three to five years. (Chock et al., 2003; See also conflicting
discussions in papers by Feichter et al., 2003; Jacobson, 2003b; see further responses by Jacobson, 2003c,
d; Penner, 2003; Penner et al., 2003).
Bond and Sun (2005) reviewed published data regarding the warming potential of BC, compared
with C02 and other GHG. Climatic effects of GHG are generally compared on the basis of
top-of-the-atmosphere, globally averaged changes in radiative balance. On that basis, BC is one of the
largest individual warming agents, after C02 and perhaps CH4 (Bond and Sun, 2005; Jacobson, 2000;
Sato et al., 2003b).
Reddy and Boucher (2007) conducted an analysis that provided regional estimates of BC emissions
from fossil fuels and biofuels. These estimates indicated that East and Southeast Asia contributed over
50% of the global BC burden and its associated direct radiative forcing. Europe was found to be the
largest BC contributor in the northern latitudes. The indirect effect of BC deposition on snow was also
estimated to be highest for Europe.
To improve understanding of the role of aerosols in climate forcing, Chung and Seinfeld (2002)
estimated the global distribution of BC, primary organic particles (those directly emitted from
combustion), secondary organic particles (primary organic compounds partially oxidized in the
atmosphere), and sulfate aerosols to model the overall radiative forcing of these groups of compounds.
The model was run with the assumption that the BC particles do not combine with organic carbon or
sulfate particles (termed an external mixture), and with the assumption that the particles are represented
by a core of BC surrounded by a shell of light scattering aerosols. Modeling results suggested an overall
radiative cooling effect from aerosols ranging from -0.39 to -0.78 W/m2.
Roberts and Jones (2008) used a climate modeling approach to compare possible effects of BC on
climate warming to those attributable to emissions from greenhouse gases. Results suggested that the
warming effect from atmospheric BC aerosols may not be large relative to that from greenhouse gases. A
different modeling approach by Roeckner et al. (2006) evaluated the effects of BC and primary organic
December 2008
9-142
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
carbon on climate under two scenarios of carbonaceous aerosol emissions. In the first scenario, BC and
primary OC emissions decreased over Europe and China, but increased at lower latitudes. In the second
scenario, emissions were frozen at 2000 levels. The effects of both scenarios on mean global temperature
were found to be small, but higher aerosol emissions at low latitudes did result in atmospheric heating and
corresponding land surface cooling that led to increased precipitation and runoff in this simulation.
9.8.1.3. Precipitation
Study of BC effects in tropical climates was undertaken by Wang (2007). Substantial effects of
direct radiative forcing by BC on the tropical Pacific were shown in model results that were similar to the
El Nino Southern Oscillation activities both in the nature and scale of effects with enhancement of the
Indian monsoon circulation. The model suggested that atmospheric heating by radiation absorption by BC
can form temperature and pressure anomalies that favor propagation of convection from western to
central and eastern Pacific. More work will be needed to distinguish between the aerosol signal and
natural factors in controlling tropical precipitation in this region.
9.8.1.4. Magnitude of Overall Direct Effects
Satellite estimates of the magnitude of the solar direct radiative effects (DRE) include effects due to
both natural aerosols and anthropogenic aerosols. In recent years, satellite estimates of the global
clear-sky DRE over oceans have advanced due to development of improved aerosol instrumentation and
algorithms (Forster et al., 2007; Penner, 2001; Yu, 2006). Kaufman et al. (2005) estimated the
anthropogenic component of the clear sky RF over ocean of-1.4 W/m2. Christopher et al. (2006)
estimated an identical number using a combination of the Moderate Resolution Imaging Spectrometer
(MODIS) instrument and the Clouds and the Earth's Radiant Energy System (CERES) broadband TOA
fluxes.
There have also been significant recent advancements in development and deployment of
surface-based remote sensing sun-photometer sites such as AERONET (Holben et al., 1998) and networks
of aerosol lidar-systems (Matthias, 2004; Murayama, 2001; Welton et al., 2001). A climatology of the
aerosol DRE was developed by Zhou et al. (2005) based on the AERONET data.
There have been recent advances in modeling the aerosol direct effect, with development of more
complete aerosol modules and their inclusion in a larger number of global atmospheric models (Forster et
al., 2007). The more complex models like those of Adams (2002), Easter (2004) and Stier (2005) now
include dynamic aerosol size distributions and changes over aerosol lifetime, and also consider mixing of
December 2008
9-143
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
the aerosol components in a more physically realistic way than was available at the time of the Third
IPCC Assessment (Forster et al., 2007).
Progress has also been made in comparing aerosol model simulations. These include the Global
Aerosol Model Intercomparison (AeroCom) initiative (Kinne, 2006; Schulz et al., 2004; Textor, 2006).
These intercomparisons help to illustrate the aspects of aerosol dynamics that are poorly constrained in
the models. For example, coarse aerosol fractions are largely responsible for variation in natural aerosol
emissions (e.g., dust and sea salt). Source strength is also strongly dependent on wind speed, adding to the
complexity of modeling natural aerosols (Forster et al., 2007). As a result, model estimates of dust
emissions for the same time period can vary by more than a factor of two depending on dust
parameterization values in the models (Balkanski et al., 2003; Luo et al., 2003; Timmreck and Shulz,
2004; Zender et al., 2004).
Global estimates of aerosol direct RF were summarized by Forster et al. (2007) as global annual
mean values at TO A, inclusive of the effects of clouds. Individual estimates, and their uncertainties, were
provided for each of the major aerosol components. Current estimates, as summarized by Forster et al.
(2007) are:
¦ sulfate, -0.4 (± 0.2) W/m2
¦ organic C from fossil fuels, -0.05 (± 0.05) W/m2
¦ black C from fossil fuels, +0.20 (± 0.15) W/m2
¦ biomass burning, +0.03 (±0.12) W/m2
¦ nitrate, -0.10 (± 0.10) W/m2
¦ mineral dust, -0.1 (± 0.2) W/m2
Forster et al. (2007) presented a combined model-based estimate of the cumulative aerosol direct
RF from all components based on multiple models and adding separate estimates for nitrate and
anthropogenic mineral dust (which are missing from most global model simulations). The overall
model-derived aerosol direct RF was estimated as -0.4 W/m2, with a 90% confidence interval of 0 to -0.8
W/m2. Three satellite-based measurement estimates suggest a stronger aerosol direct RF than the model
simulations of approximately -0.55 W/m2 (Bellouin et al., 2005; Chung et al., 2005; Yu, 2006).
December 2008
9-144
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
9.8.2. Indirect Effects
Aerosols can interact with clouds and precipitation in a variety of ways with processes outlined by
IPCC (Denman et al., 2007) and summarized in Tables 9-5 and 9-6. Such cloud feedbacks remain the
largest source of uncertainty in climate estimates.
Because clouds form with the aid of hygroscopic aerosols, anthropogenic PM can have indirect
effects on climate by altering cloud microphysical processes. Indirect effects of aerosols on climate relate
in part to perturbation of the albedo of clouds. The first indirect effect from increased cloud condensation
nuclei refers to increased cloud droplet concentrations, smaller droplet radii, and more reflective clouds.
The second indirect effect refers to the observation that decreased cloud droplet effective radii can cause
lower coalescence rates, reduced precipitation, longer cloud lifetime, and greater cloud spatial extent
(Kanakidou et al., 2005). Even small changes in cloud properties over global scales can have substantial
effects on the amount of solar radiation absorbed by the Earth, and therefore have an important effect on
climate. Much of the uncertainty associated with predicting effects on cloud radiative properties is due to
variability in cloud droplet number as influenced by aerosol characteristics. Model simulations reported
by Kristjansson (2002) suggested that BC PM contributes only marginally to these indirect effects.
Table 9-5.
Overview of the different aerosol indirect effects and their sign of the net radiative flux
change at the top of the atmosphere (TOA).
Effect
Cloud Types
Affected
Process
Sign of Change
in TOA
Radiation
Potential
Magnitude
Scientific
Understanding
Cloud albedo
effect
All clouds
For the same cloud water or ice content more but
smaller cloud particles reflect more solar radiation
Negative
Medium
Low
Cloud lifetime
effect
All clouds
Smaller cloud particles decrease the precipitation
efficiency thereby presumably prolonging cloud
lifetime
Negative
Medium
Very low
Semi-direct effect
All clouds
Absorption of solar radiation by absorbing
aerosols affects static stability and the surface
energy budget, and may lead to an evaporation of
cloud particles
Positive or
Negative
Small
Very low
Glaciation indirect
effect
Mixed-phase
clouds
An increase in IN increases the precipitation
efficiency
Positive
Medium
Very low
Thermodynamic
effect
Mixed-phase
clouds
Smaller cloud droplets delay freezing causing
super-cooled clouds to extend to colder
temperatures
Positive or
Negative
Medium
Very low
Source: Denman (2007)
December 2008
9-145
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Table 9-6. Overview of the different aerosol indirect effects and their implications for the global
mean net shortwave radiation of the surface Fsfc (columns 2-4) and for precipitation
(columns 5-7).
Effect
Sign of
Change in
Fsfc
Potential
Magnitude
Scientific
Understanding
Sign of Change in
Precipitation
Potential
Magnitude
Scientific
Understanding
Cloud albedo
effect
Negative
Medium
Low
Cloud lifetime
effect
Negative
Medium
Very low
Negative
Small
Very low
Semi-direct effect Negative
Large
Very low
Negative
Large
Very low
Glaciation indirect Positive
effect
Medium
Very low
Positive
Medium
Very low
Thermodynamic
effect
Positive or
Negative
Medium
Very low
Positive or Negative
Medium
Very low
Source: Denman (2007)
9.8.2.1. First Indirect Effect: Cloud Albedo
The best characterized indirect effects included in climate models is the increase in cloud droplets
(Lohman and Feichter, 2005). This occurs if the aerosol concentration increases without a change in the
moisture content of the cloud. The size and number of droplets decreases, thereby increasing the albedo of
the cloud and the reflection of radiation back into space. This process also can suppress precipitation,
which increases cloud life and reduces aerosol washout.
The effects of PM on cloud formation and cloud processes are magnified by the multitude of
linkages between clouds and the overall climate system. In particular, clouds affect solar and terrestrial
radiation and precipitation formation. Interactions between aerosols and clouds are complex, sometimes
non-linear, and not well quantified (Ramaswamy et al., 2001). On average, clouds have a cooling effects
on the present climate of the earth (Quante, 2004); however, a small change in the amount and
distribution of cloud cover could have an important effect on the total energy budget of the planet.
Estimates of current simulations of the magnitude of the cloud albedo effect were summarized by
Forster et al. (2007). Increased concentrations of CCN or ice nuclei (IN) due to human activities can
modify the properties of clouds and affect climate by increasing the albedo of clouds (Jacob et al., 2005;
Penner, 2001; Ramanathan et al., 2001).
December 2008
9-146
DRAFT-DO NOT QUOTE OR CITE
-------
Radiative forcing of climate between 1750 and 2005
Radiative Forcing Terms Climate efficacy Spatial scale
1
,
Long-lived f
;
(see
caption)
Global
High
N?0 I
nraonhmico naQac 1
-10-
100 yrs
I
!
IK
1.0-1.2
Global
High
:
• l.y|rjcar::::rs
Ozone
Stratospheric l-j
' (-0,05) 1
h—| Tropospheric
0.5-2/)
Weeks to
100 yrs
Continental
togbbal
Med
Stratospheric water
vapour from CH4
!
^i '
~1.0
10 years
Global
Low
Surface albedo
Lard use ' M
1 Black carbon
^ on snow
—
10-
100 yrs
Local to
continental
Med
Total
' Direct effect
1—9
0.7-1.1
Days
Conlinenlal
to global
Med
Aerosol
Clojd albcco
i |—
|
1.0-2X1
Hours -
Days
Continental
to global
L°»
Linear contrails
(0.01) !
-0.6
Hours
Continental
Low
Solar irradiance
:
K ! .
0.7-IX)
10-
100 yrs
Global
l™
Radiative Forcing (W m"2)
2
£
!5
(0
n
o
CL
J3
c
tr
0
-3-2-101234
Radiative Forcing (W m"2)
Source: Forsteret al. (2007)
Figure 9-57. (A) Global mean RFs from the agents and mechanisms discussed in Forster et al.
(2007) grouped by agent type. Anthropogenic RFs and the natural direct solar RF are
shown. Columns indicate other characteristics of the RF; efficacies are not used to
modify the RFs shown. Time scales represent the length of time that a given RF term
would persist in the atmosphere after the associated emissions and changes ceased.
No CO2 time scale is given, as its removal from the atmosphere involves a range of
processes that can span long time scales, and thus cannot be expressed accurately
with a narrow range of lifetime values. (B) Probability distribution functions (PDFs)
generated by combining human-caused RFs given above in panel A. Three cases are
shown: the total of all anthropogenic RF terms (block filled red curve); LLGHGs and
ozone RFs only (dashed red curve); and aerosol direct and cloud albedo RFs only
(dashed blue curve). Surface albedo, contrails and stratospheric water vapor RFs are
included in the total curve but not in the others. For all of the contributing forcing
agents, the uncertainty is assumed to be represented by a normal distribution (and
90% confidence intervals) with the following exceptions: contrails, for which a
lognormal distribution is assumed to account for the fact that the uncertainty is
quoted as a factor of three; and tropospheric ozone, the direct aerosol RF (sulfate,
fossil fuel organic and BC, biomass burning aerosols) and the cloud albedo RF, for
which discrete values are randomly sampled. Additional normal distributions are
included in the direct aerosol effect for nitrate and mineral dust. A one-million point
Monte Carlo simulation was performed to derive the PDFs (Boucher and Haywood,
2001). Natural RFs (solar and volcanic) are not included in these three PDFs. Climate
efficacies are not accounted for in forming the PDFs.
warning
Long-lived greenhouse
and ozone
radiative forcings
Total aerosol
radiative forcing , -
Total anthropogenic radiative forcing
December 2008
9-147
DRAFT-DO NOT QUOTE OR CITE
-------
Components of radiative forcing for principal emissions
0,(T) H,0(S
Ha ocarbons
CFCr. HCFCs nalrns
HFCs
-HfCs
N rata
WVOC
Black carbori
Si.lfatfi
(d'rect)
I
Black carbon
fsnow albedo)
Organic carbon
(direct)
Mine'al dus1
Cloud alo&do eftecl
Contrails
Surface albedo
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
sizes (Andreae et al., 2004; Mircea, 2005). The ability of a given particle to act as effective CCN may be
more strongly controlled by the size of the particle, rather than particle composition (Dusek et al., 2006).
Numerous difficulties remain, however, in quantifying such effects (cf. Forster et al., 2007; McFiggans
and Midgley, 2001).
Atmospheric models estimate a negative global mean RF due to the cloud albedo effect from PM,
with substantial variation among the estimates of the magnitude of this effect, from -0.22 to -1.85 W/m2
(Forster et al., 2007). Variation in these estimates stems largely from differences among model treatment
of aerosol, cloud processes, and aerosol-cloud interaction processes (Forster et al., 2007). The effect is
generally larger over land than oceans, and the model estimates are more variable (and uncertain) over
land (Lohman and Feichter, 2005). Forster et al. (2007) provided a best estimate for the cloud albedo RF
of -0.7 W/m2, with a 90% confidence range (5%-95% uncertainty range) of -0.3 to -1.8 W/m2.
The RFs discussed by Forster et al. (cf. 2007) and summarized here are illustrated in Figure 9-57
along with their uncertainty ranges. This helps put the estimates of RF from aerosol particles into the
context of RF estimates from long lived greenhouse gasses and other agents of warming. The combined
aerosol direct effect and cloud albedo indirect effect exert an RF that is estimated to be about -1.3 W/m2,
with a 90% confidence range of -22 to -0.5 W/m2. A breakdown of the estimated RF for each of the
principal gas- and aerosol-phase emissions components is given in Figure 9-58. The estimated indirect
cloud albedo effect is larger than any of the individual component direct effects. Among the aerosol
components, only BC is believed to exert a positive RF.
Precipitation suppression by PM has been known for several decades and has been quantified in
recent studies. It is believed to result primarily from the cloud albedo indirect effect. Increases in CCN
concentrations downwind of field burning sites were investigated by Warner and Twomey (1967) and
Warner (1968). Albrecht (1989) hypothesized that higher water droplet concentrations in ship-track trails,
due to the high CCN concentration and consequent formation of more small cloud droplets, would reduce
drizzle drop formation. Aircraft observations showed that ship-track clouds do actually contain higher
cloud droplet concentrations than surrounding clouds, smaller droplet sizes, and only 10% of the number
of drizzle drops as the surrounding clouds (Radke et al., 1989). Rosenfeld (1999, 2000) identified
pollution tracks downwind of pollution sources comprised of highly reflective cloud droplets. These
pollution-track clouds have smaller effective cloud particle size than clouds in surrounding less polluted
areas, resulting in precipitation suppression. More recently, Borys et al. (2003) found that winter
orographic precipitation can be reduced by up to 50% by decreased riming efficiency of ice crystals when
pollution aerosols contribute to smaller cloud droplets.
The effects of aerosols on cloud processes over land have been shown to vary compared with those
over the ocean. Rosenfeld et al. (2002) found that increased concentrations of aerosol pollutants over the
December 2008
9-149
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
ocean suppressed precipitation to a much lesser degree than over land. This difference was explained by
the tendency of sea salt nuclei to initiate formation of droplets that can readily combine with pollutant
particles. This process also results in a lower atmospheric residence time for pollutant aerosols over
oceans than over land, thus contributing to the cloud lifetime effect described below.
9.8.2.2. Second Indirect Effect: Cloud Lifetime
As discussed by Denman et al. (2008), the cloud albedo effect cannot be easily separated from
other effects, especially from the cloud lifetime effect. The processes that decrease cloud droplet size per
given liquid water content also decrease the formation of precipitation, and therefore presumably prolong
cloud lifetime. An increase in cloud lifetime would be expected to feed back to cloud albedo.
To estimate or model the radiative effects of aerosols, it is important to determine the dependence
of aerosol optical properties on relative humidity (RH). Such information is needed to quantify the
influence of atmospheric aerosols on climate. Relative humidity controls the water content of atmospheric
particles, and the magnitude of the RH influence depends on aerosol size and composition (Baynard et al.,
2006). Water uptake properties of particulate OM must also be simplified and approximated (Kanakidou
et al., 2005; Quinn et al., 2005) in order to more fully characterize the RH dependence of light scattering
by PM.
Representations of aerosol-cloud and convection-cloud interactions in climate models are crude
(Lohman and Feichter, 2005). Cloud cover is variable and inhomogeneous (2007). Models do not
consistently provide accurate estimates of the amounts of cloud liquid and ice water content, and GCMs
do not resolve the small scales at which aerosol-cloud interactions occur. In addition, differences in the
horizontal and vertical resolution of cloud occurrence limit accurate representation of the shallow warm
cloud layers over the oceans most susceptible to changes due to anthropogenic aerosol particles (2007).
One important aspect of the cloud albedo effect involves effects on local and regional precipitation
budgets. The suppression of precipitation by PM pollution is supported by a large body of recent research.
Small atmospheric particles slow precipitation formation in clouds. For orographic clouds, which are
shallow and short-lived, there is a net decrease in precipitation amount. As a further consequence of the
decrease in precipitation, cloud lifetime is extended, with additional feedback to albedo and other climate
impacts. Increased cloud lifetime and extent results in a further increase in the reflection of solar radiation
(Ramanathan et al., 2001).
The complexities of aerosol effects on cloud processes under varying relative humidity were
investigated by Fan et al. (2008). This study showed that aerosol effects on cloud microphysical
properties and precipitation were negligible at a relative humidity of 40%, but were significant at a
relative humidity of 60-70%. Effects were also shown to vary between continental clouds and marine
December 2008
9-150
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
clouds, as well as with composition and concentration of aerosols. Additional information on aerosol
effects on meteorological relationships were shown by Massie et al. (2007), who reported that aerosol
indirect effects were a function of the atmospheric pressure at the top of clouds. This information was
useful in improving model results that simulate increases in cloud reflectance from non absorbing
aerosols.
Atmospheric soot particles up to 10 ^m can influence climate processes by altering the radiation
balance through cloud formation. In addition, soot particles can act as condensation nuclei to change
precipitation patterns ((Hansen and Nazarenko, 2004; Sato et al., 2003b).
The effects of low solubility organic aerosols on droplet formation rates relative to levels of
inorganic pollutant aerosols were investigated by Shantz et al. (2003). Less soluble organic acid particles
were found to delay droplet formation relative to more soluble (NH^SO^ Results also suggested that
internal mixing of relatively insoluble organic compounds with (NH^SC^ also delayed droplet formation
relative to the inorganic particle. Overall delay in droplet formation from low-solubility organics was
estimated to reduce droplet numbers in clouds by up to 85%. Particle solubility was indicated as a
potentially important factor in the precipitation efficiency of clouds.
One aspect of carbonaceous aerosol is its capacity to provide CCN. These PM constituents were
previously thought to be hydrophobic and therefore not involved in cloud formation. More recent research
has suggested that these particles are largely hydrophobic at the time of emission, but that they acquire
hydrophilic characteristics with ageing (Alves et al., 2006). During long-range transport, the organic
compounds in the aerosol phase are oxidized into secondary products and largely transfer to the
particulate phase that is much less volatile and more water soluble (Atkinson, 1994). At remote sites, the
organic aerosol may be the main source of CCN; even in less remote regions, one-third to one-half of the
total organic mass can be water soluble (Alves et al., 2006).
More comprehensive modeling efforts have recently added formulations to combine direct
radiative absorption with the indirect climate forcing from effects of aerosols on cloud processes in
climate models. Kristjansson et al. (2005) modeled the relative effects of direct and indirect radiative
forcing by a suite of aerosols that are prevalent in the atmosphere. This modeling exercise indicated that
the indirect cooling effect was much stronger than the warming effect from direct absorption, but that the
uncertainly was large. In fact, these authors stated that the uncertainly was so large that it could be
possible that indirect cooling is presently offsetting much of the current C02 warming effect. Jacobson
(2006) modeled emissions and effects of BC on both direct radiative forcing and indirect effects of
hydrometeor particle formation (cloud droplets that incorporate BC) to determine the climate effect of
BC. The heat generated from radiative absorption by BC within the cloud was found to increase water
vapor and decrease precipitation. The increase in water vapor at the expense of precipitation contributed
December 2008
9-151
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
to warming beyond that caused by BC absorption within clouds. This result contrasts with the increase in
albedo and associated cooling attributed to increased concentrations of non-absorbing aerosols that
decrease droplet size and increase the reflectiveness of clouds.
First-principle approaches have been developed in which cloud droplet number is computed for
each GCM grid cell (e.g., Jacobson, 2003a, 2004). The indirect effects of carbonaceous aerosols are
believed to be of only minor importance (Jacobson, 2003c).
Uncertainty in model results stems in part from a lack of atmospheric measurements. Recently,
Kaufman (2006) was able to produce the first measurement-based estimates of the anthropogenic aerosol
fraction and of the impact of aerosol on the cloud cover and height with MODIS (Moderate Resolution
Imaging Spectroradiometer) satellite imagery. Further work of this type is necessary to evaluate the wide
range of modeling results.
9.8.3. Other Effects
There are several other kinds of climate effects from aerosol PM. None is well understood or well
quantified. The semi-direct effect, which involves absorption of solar radiation by soot particles followed
by re-emission as thermal radiation, is expected to heat the air mass and increase its static stability relative
to the surface. The semi-direct effect can also cause evaporation of cloud droplets, thereby partially
offsetting the cloud albedo indirect effect. The glaciation effect involves an increase in IN, which is
expected to cause rapid glaciation of a super-cooled liquid water cloud due to the differences in vapor
pressure over ice and water. Unlike cloud droplets, these ice crystals can quickly reach precipitation size,
with the potential to turn a non-precipitating cloud into a precipitating cloud. The thermodynamic effect
involves a delay in freezing by the smaller cloud droplets, which can cause super cooled clouds to occur
under colder temperatures. The possible consequences to radiative flux of all of the processes are outlined
in Table 9-5 (top of the atmosphere effects) and Table 9-6 (surface radiative and precipitation effects;
Denman et al., 2007), though significant uncertainties remain. Nevertheless, the individual processes
cannot be considered in isolation because of the numerous feedbacks, and because atmospheric aerosol
concentrations and climate are intimately coupled (Denman et al., 2007; Dentener, 2006).
9.8.4. Effects on Local and Regional Climate
Most effects of PM on climate, as assessed by IPCC (c.fi, Stohl et al., 2007) and summarized in this
assessment, focus on global-scale processes and responses. In addition, it is also possible that PM
emissions contribute to local and regional climate changes. These might include short-term cycles in
December 2008
9-152
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
rainfall or temperature and rainfall suppression, especially near cities and for orographic precipitation.
Rainfall suppression, in particular, is believed to exacerbate water supply problems which are substantial
in many regions, especially in the western U.S.
Aerosol particles, directly and through cloud enhancement, may reduce near-surface wind speeds
locally. Slower winds, in turn, reduce evaporation. The overall impact can be a reduction in local
precipitation. Jacobson and Kaufman (2006) investigated the effects of PM on spatially-distributed wind
speeds and resulting feedbacks to precipitation using the GATOR-GCMOM (Jacobson, 2001) and
supporting evidence from satellite data. The study focused on the South Coast Air Basin (SCAB) in
California during February and August, 2002-2004. The modeled precipitation decrease over land in
California was 2% of the baseline 1.5 mm/day due to emissions of anthropogenic aerosol particle and
precursor gasses in the SCAB domain. However, the reduction over much of the Sierra Nevada, where
most precipitation falls, was up to 0.5 mm/day, or 4 to 5% of the baseline 10 to 13 mm/day in that
mountainous region (Jacobson and Kaufman, 2006). The probable mechanism was described as follows.
Aerosol particles and aerosol-enhanced clouds reduce wind speeds below them by stabilizing the air,
reducing the vertical transport of horizontal momentum. In turn, the reduced wind speeds, and associated
reduced evaporation and increased cloud lifetime, contributes to reduced local and regional precipitation
(Jacobson and Kaufman, 2006).
Effects of air pollution on regional precipitation were quantified by Givati and Rosenfeld (2004).
They found a 15 to 25% reduction in the orographic component of precipitation downwind of major
coastal urban areas during the 20th century. Their study focused on orographically-forced clouds because
these short-lived, shallow clouds are expected to exhibit the largest effect of air pollution on precipitation.
Substantially larger precipitation suppression due to aerosol particulate pollution was found between
Fresno and Sacramento in California by Givati (2004). Precipitation losses over topographical barriers in
the Sierra Nevada amounted to 15%-25% of the annual precipitation at elevations less than 2,000 m. This
precipitation suppression occurred mainly in the relatively shallow orographic clouds within the cold air
mass of cyclones. The suppression that occurred on the upslope side of the mountains was coupled with
similar percentage (but lower absolute volume) enhancement on the drier downslope eastern side (Givati
and Rosenfeld, 2004). Similar results were found in studies by Griffith et al. (2005), Jirak and Cotton
(Jirak and Cotton, 2006), Rosenfeld and Givati (2006), and Rosenfeld et al. (2007). At all of these study
locations (California, Israel, Utah, Colorado, China), orographic precipitation decreased by 15 to 30%
downwind of pollution sources, likely due to creation of more and smaller cloud droplets and resulting
suppression of precipitation.
The study of Givati and Rosenfeld (2004) was the first to quantify the microphysical effect of
mesoscale precipitation. Following the findings of Givati and Rosenfeld (2004), the effects of aerosol air
December 2008
9-153
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
pollution on precipitation at high elevation sites in the Front Range of Colorado adjacent to urban areas
were investigated by Jirak and Cotton (2006). Examination of precipitation trends showed that the ratio of
upslope precipitation during easterly flows at high elevation west of Denver and Colorado Springs to the
upwind urban sites decreased by about 30% over the past half century. These results provide further
support for the hypothesis that aerosol pollution suppresses orographic precipitation downwind of
pollution source areas.
Griffith et al. (2005) found similar reductions in mountainous precipitation in Utah, downwind of
Salt Lake City and Provo. The ratio of precipitation at mountain stations located in rural settings in Utah
and Nevada remained stable, supporting the hypothesis that air pollution decreases Ro (the ratio of
precipitation at the downwind site to precipitation at the upwind pollution source) over the mountains to
the east of Salt Lake City.
Rosenfeld and Givati (2006) extended the investigation of the suppression of precipitation by
aerosol pollutants to a larger scale by examining the ratio between precipitation amounts over the hills to
precipitation over upwind lowland areas throughout the western U.S. from the Pacific Coast to the Rocky
Mountains. They found in these paired analyses a pattern of decreasing precipitation by as much as 24%
from the Mexican border to central California, with no decrease in northern California and Oregon and
smaller decrease of 14% in Washington east of Seattle and Puget Sound. Similar decreases were found
over Arizona and New Mexico (Rosenfeld and Givati, 2006), Utah (Griffith et al., 2005), and the east
slope of the Colorado Rockies (Jirak and Cotton, 2006).
Suppression of winter orographic precipitation appears to occur up to hundreds of kilometers
inland of coastal urban areas (Rosenfeld and Givati, 2006). Decreases in this precipitation ratio occurred
during winter orographic precipitation, but not during convective summer precipitation over the same
mountain ranges. This finding agrees with the expectation that aerosol-induced changes in the rate of
precipitation formation would cause a decrease in precipitation from shallow and short-lived orographic
clouds, but not necessarily from deeper and longer-lived thermally-driven convective clouds.
Results of these studies of aerosol effects on orographic precipitation suggest that human-caused
air pollution, and fine particles in particular, have had a large effect on precipitation well beyond the local
scales of the pollution sources (Rosenfeld and Givati, 2006).
9.8.5. Glaciers and Snowpack
Organic compounds are incorporated into snow by wet and dry deposition processes (Lei and
Wania, 2004; Roth et al., 2004). Atmospherically deposited organics appear to be ubiquitous in
snowpacks at appreciable concentrations (Grannas et al., 2007). Examples include PAHs, phthalates,
December 2008
9-154
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
alkanes, phenols, low molecular weight carbonyls, POPs, and low molecular weight organic acids (cf.,
Halsall, 2004; Nakamura et al., 2000; Villa et al., 2003). Humic-like substances found in the snowpack
may release VOCs into the atmosphere via photo-oxidation (Grannas et al., 2004; Grannas et al., 2007).
Several thousand organic species were identified by Grannas et al. (2006), based on molecular weight,
from a single ice core collected in Russia. Little information is available, however, regarding the chemical
properties of these chemical constituents. In addition to the diversity of chemicals that are deposited into
the snowpack, there are also biological organisms, including bacteria and algae. Their role in influencing
snow chemistry and volatilization processes are not understood (Grannas et al., 2007).
Recent research has explored connections between the atmosphere and the cryosphere (land or sea
covered by snow or ice). A seasonal maximum of 40% of the Earth's land surface is covered by snow or
ice, as well as several percent of the oceans. Particulate deposition to snow and ice surfaces can affect
melting rates. Deposition of PM to glacial ice surfaces can affect the subsequent rate of melting. A thin
cover of debris contributes to accelerated melting. A thicker cover of debris, such as may result from a
volcanic eruption, retards melting. The difference is due to the changing balance between enhanced
absorption of shortwave radiation by PM and conductive heat flow (insulation) through a buildup of
material having low heat conduction (Kirkbride and Dugmore, 2003).This issue is particularly important
for deposition of large quantities of volcanic material. To a lesser extent, however, the same principles
apply to PM deposition derived from air pollution. Under a thin layer of debris, ablation rates are higher
than for clean ice. However, as the thickness of the debris layer increases, ablation rates systematically
decline (Nicholson and Benn, 2006). The threshold debris thickness that separates ablation increase from
decrease is site specific and depends on local climate and the nature of the debris particles. Nicholson and
Benn (2006) presented a surface energy balance model to calculate ice melt beneath a surface debris
layer, based on meteorological data and basic debris characteristics. Modeled melting rates matched
observed rates, suggesting that the model produced useful results.
Long-range atmospheric transport of PM delivers a large fraction of the total input of POPs to the
Arctic region (Halsall, 2004). These contaminants can accumulate in Arctic food webs and have become
the focus of international research and concern. Nevertheless, fate and transport of POPs within terrestrial
and marine Arctic ecosystems are not well understood and are strongly affected by the presence of snow
and ice. Sea ice provides a barrier to air-water exchange, and this hinders volatilization and re-emission of
previously deposited contaminants (Halsall, 2004). Thus, the effects of greenhouse gasses and PM on
climate in the Arctic region have feedbacks to POP fate, transport, and toxicity. The transfer of POPs
among the major abiotic environmental compartments in the Arctic are summarized in Figure 9-59 from
Halsall (2004). Recent studies detailing rate and transport of POPs are summarized in Table 9-7.
December 2008
9-155
DRAFT-DO NOT QUOTE OR CITE
-------
Table 9-7. Recent studies highlighting POP occurrence and fate in the major arctic compartments.
Atmosphere
1
Annual time-series of OC and PCB concentrations in the Norwegian Arctic
Oehme et al. (1996)
2
Long-term analysis of the chlordane-group and their input to the Arctic with changing sources
Bidleman et al. (2002)
3
PAH occurrence at monitoring sites across the Arctic, seasonality and gas/particle partitioning
Halsall et al. (1997)
4
PCB occurrence at monitoring sites across the Arctic, spatial differences and seasonality
Stern et al. (1997)
5
Long-term analysis of PCB and OC trends in the Canadian Arctic and seasonal patterns
Hung et al. (2001; 2002)
6
T rans-Pacific LRAT and impact of Asian sources on the western Canadian Arctic
Bailey et al. (2000)
FRESHWATER
7
Annual average water concentrations in major Russian rivers for selected OC pesticides
Alexeeva et al. (2001)
8
Long-term (decades) PCB deposition profile in Arctic lake sediments
Muiret al. (1996)
9
Mass balance of selected OCs in Canadian Arctic lake conducted with data collected over 3 years
Helm et al. (2002)
10
Examining the biodegradation of HCHs in Canadian Arctic watersheds
Helm et al. (2000)
MARINE
11
Transport and entry of (3-HCH into western Arctic Ocean via Pacific surface waters
Li et al. (2002)
12
Occurrence of current use pesticides in air, fog and surface seawater in the
western Arctic Ocean
Chermyak et al. (1996)
13
Resolving petrogenic and anthropogenic PAH input to marine sediments in coastal Arctic seas
Yunker et al. (1996)
14
Quantifying abiotic and biotic degradation of HCHs in the Arctic Ocean water column
Harner et al. (2000)
15
PCBs and OCs in surface ocean water—Bering and Chukchi seas
Strachan et al. (2001)
16
Spatial patterns of HCHs and toxaphene in Arctic Ocean surface water
Jantunen and Bidleman (1998)
SNOW/AIR-FRESHWATER
17
PAHs (and inorganics) in surface snow layers (snowpit) at Summit, Greenland
Masclet et al. (2000)
18
PAHs measured in snow and firn layers on Agassiz ice-cap, Ellesmere Island, Canada
Peters et al. (1995)
19
Modelling OC behaviour and fate in the surface seasonal snow pack at
Amituk Lake, Canada
Wania et al. (1998)
20
OCs, PCBs and PAHs in snow and ice of the Ob-Yenisey watershed of the Russian Arctic
Melnikov et al. (2003)
OCEAN/AIR
21
Transfer of a-HCH across the air/water interface in the western Arctic ocean
Jantunen and Bidleman (1996)
22
Calculated seasonality of OC air/water fluxes in the Canadian high Arctic
Hargrave et al. (1997)
OCEAN/ICE
23
Transport potential of contaminants across the Arctic ocean via sea-ice drift
Pfirman et al. (1997)
24
The importance of eastern Arctic sea-ice drift as a source of contaminants to the Norwegian sea
Korsnes et al. (Korsnes et al., 2002)
Source: Halsall (2004)
December 2008
9-156
DRAFT-DO NOT QUOTE OR CITE
-------
LRA1
Atmosphere
""H 9
(M)
(19)
A
V
(21-22
A
> f
— : fTT-W
Snow/icc caps
I v/i iv~ ;i :;i
I .ilvhirvr.: 11
SI
Snow.sea too
rivers
" ,1 ROT
{H II., _ ->
Surface
Ocean
Deep Ocean
Figure 9-59. The transfer of POPs between the major abiotic compartments of the Arctic. Shaded
arrows represent inputs/outputs of POPs to the Arctic. The numbers refer to selected
studies detailing the occurrence and behavior of POPs, and are listed in Table 9-7.
Question marks represent those areas that are least well understood. LRAT-long
range atmospheric transport; LROT - long range oceanic transport.
9.8.6. Global Warming Potentials
1 One approach to making comparisons of effects from the many and varied contributors to climate
2 warming involves use of global warming potentials (GWPs) defined as the integral of the radiative
3 forcing caused by the pulse emission of 1 kg of a chemical species of a time horizon T, which results in a
4 unit of W m2/kg/yr (Boucher and Reddy, 2008). It depends on both the radiative efficiency of the
5 chemical species and its decay time.
6 Bond and Sun (2005) estimated that BC has a high GWP of 680 even though it has a short lifetime
7 in the atmosphere. Reddy and Boucher (2007) found that the direct GWP for BC depends on its source
8 region. Such regional differences reflect differences in atmospheric lifetime, which are largely due to the
9 regional efficacy of wet deposition as a process that removes PM from the atmosphere. The GWP of BC
10 is still highly uncertain (Boucher and Reddy, 2008).
December 2008
9-157
DRAFT-DO NOT QUOTE OR CITE
-------
References
Aam BB; Fonnum F. (2007). ROS scavenging effects of organic extract of diesel exhaust particles on human neutrophil
granulocytes and rat alveolar macrophages. Toxicology 230: 207-218.
Abbey DE; Lebowitz MD; Mills PK; Petersen FF; Beeson WL; Burchette RJ. (1995). Long-term ambient concentrations of
particulates and oxidants and development of chronic disease in a cohort of nonsmoking California residents. Inhalation
Toxicology 7: 19-34.
Abbey DE; Nishino N; McDonnell WF; Burchette RJ; Knutsen SF; Lawrence Beeson W; Yang JX. (1999). Long-term inhalable
particles and other air pollutants related to mortality in nonsmokers. American journal of respiratory and critical care
medicine 159: 373-382.
Aboal JR; Fernandez JA; Carballeira A. (2004). Oak leaves and pine needles as biomonitors of airborne trace elements pollution.
Environmental and Experimental Botany 51:215-225.
Abou Chakra OR; Joyeux M; Nerriere E; Strub MP; Zmirou-Navier D. (2007). Genotoxicity of organic extracts of urban airborne
particulate matter: an assessment within a personal exposure study. Chemosphere 66: 1375-1381.
ABT. (2001). Assessing Public Opinions on Visibility Impairment Due to Air Pollution: Summary Report. Abt Associates, Inc.
Prepared for EPA Office of Air Quality Planning and Standards; funded under EPA Contract No. 68-D-98-001. Bethesda,
Maryland. January 2001. Available at http://www.epa.gov/ttncaaal/tl/reports/vis_rpt_final.pdf. Accessed 9/16/2008,
ABT. (2002). Sense of Place and Stewardship: Final Focus Group Report. Abt Associates, Inc. Prepared for Southern
Appalachian Mountains Initiative Asheville, NC. Bethesda, Maryland,
ACGIH. (2005). TLVs and BEIs: Based on the documentation of the threshold limit values for chemical substances and physical
agents and biological exposure indices. Cincinnati, OH: American Conference of Governmental Industrial Hygienists
Worldwide.
Achenbach S; Daniel WG. (2001). Noninvasive coronary angiography—an acceptable alternative? The New England journal of
medicine 345: 1909-1910.
Ackermann-Liebrich U; Kuna-Dibbert B; Probst-Hensch NM; Schindler C; Felber Dietrich D; Stutz EZ; Bayer-Oglesby L; Baum
F; Brandli O; Brutsche M; Downs SH; Keidel D; Gerbase MW; Imboden M; Keller R; Knopfli B; Kunzli N; Nicod L;
Pons M; Staedele P; Tschopp JM; Zellweger JP; Leuenberger P. (2005). Follow-up of the Swiss Cohort Study on Air
Pollution and Lung Diseases in Adults (SAPALDIA 2) 1991-2003: methods and characterization of participants. Soz
Praventivmed 50: 245-263.
Ackermann-Liebrich U; Leuenberger P; Schwartz J; Schindler C; Monn C; Bolognini G; Bongard JP; Brandli O; Domenighetti
G; Elsasser S; Grize L; Karrer W; Keller R; Keller-Wossidlo H; Kunzli N; Martin BW; Medici TC; Perruchoud AP;
Schoni MH; Tschopp JM; Villiger B; Wuthrich B; Zellweger JP; Zemp E. (1997). Lung function and long term exposure
to air pollutants in Switzerland. Study on Air Pollution and Lung Diseases in Adults (SAPALDIA) Team. American
journal of respiratory and critical care medicine 155: 122-129.
Adachi K; Tainosho Y. (2004). Characterization of heavy metal particles embedded in tire dust. Environment International 30:
1009-1017.
Adamkiewicz G; Ebelt S; Syring M; Slater J; Speizer FE; Schwartz J; Suh H; Gold DR. (2004). Association between air pollution
exposure and exhaled nitric oxide in an elderly population. Thorax 59: 204-209.
Adamo P; Crisafulli P; Giordano S; Minganti V; Modenesi P; Monaci F; Pittao E; Tretiach M; Bargagli R. (2007). Lichen and
moss bags as monitoring devices in urban areas. Part II: Trace element content in living and dead biomonitors and
comparison with synthetic materials. Environmental Pollution 146: 392-399.
Adams HS; Kenny LC; Nieuwenhuijsen MJ; Colvile RN; Gussman RA. (2001). Design and validation of a high-flow personal
sampler for PM2.5. Journal of exposure analysis and environmental epidemiology 11: 5-11.
Adams PM; Seinfeld JH. (2002). Predicting global aerosol size distributions in general circulation models. Journal of
Geophysical Research 107(D19), 4370, doi:10.1029/2001JD001010.
Adams SM. (2003). Establishing Causality between Environmental Stressors and Effects on Aquatic Ecosystems. Human and
Ecological Risk Assessment 9: 17-35.
Adamson IY; Vincent R; Bakowska J. (2003). Differential production of metalloproteinases after instilling various urban air
particle samples to rat lung. Exp Lung Res 29: 375-388.
Adar SD; Adamkiewicz G; Gold DR; Schwartz J; Coull BA; Suh H. (2007b). Ambient and microenvironmental particles and
exhaled nitric oxide before and after a group bus trip. Environ Health Perspect 115: 507-512.
Adar SD; Gold DR; Coull BA; Schwartz J; Stone P; Suh H. (2007a). Focused exposure to airborne traffic particles and heart rate
variability in the elderly. Epidemiology 18: 95-103.
Adar SD; Kaufman JD. (2007). Cardiovascular disease and air pollutants: evaluating and improving epidemiological data
implicating traffic exposure. Inhal Toxicol 19 Suppl 1: 135-149.
Adgate JL; Mongin SJ; Pratt GC; Zhang J; Field MP; Ramachandran G; Sexton K. (2007). Relationships between personal,
indoor, and outdoor exposures to trace elements in PM(2.5). Sci Total Environ 386: 21-32.
Adgate JL; Ramachandran G; Pratt GC; Waller LA; Sexton K. (2003). Longitudinal variability in outdoor, indoor, and personal
PM2.5 exposure in healthy non-smoking adults. Atmospheric Environment 37: 993-1002.
December 2008 R-1 DRAFT-DO NOT QUOTE OR CITE
-------
Adgate JL; Waller LA; Sexton K; Ramachandran G; Pratt GC. (2002). Spatial and temporal variability in outdoor, indoor, and
personal PM2.5 exposure. Atmospheric Environment 36: 3255-3265.
Aekplakorn W; Loomis D; Vichit-VadakanN; Shy C; Plungchuchon S. (2003). Acute effects of S02 and particles from a power
plant on respiratory symptoms of children, Thailand. Southeast Asian J Trop Med Public Health 34: 906-914.
Agarwal R; Jayaraman G; Anand S; Marimuthu P. (2006). Assessing respiratory morbidity through pollution status and
meteorological conditions for Delhi. Environ Monit Assess 114: 489-504.
Agatston AS; Janowitz WR; Hildner FJ; ZusmerNR; Viamonte M, Jr.; Detrano R. (1990). Quantification of coronary artery
calcium using ultrafast computed tomography. J Am Coll Cardiol 15: 827-832.
Agopyan N; Bhatti T; Yu S; Simon SA. (2003). Vanilloid receptor activation by 2-and 10-(im particles induces responses leading
to apoptosis in human airway epithelial cells. Toxicology and applied pharmacology 192: 21-35.
Ahsan MK; Nakamura H; Tanito M; Yamada K; Utsumi H; Yodoi J. (2005). Thioredoxin-1 suppresses lung injury and apoptosis
induced by diesel exhaust particles (DEP) by scavenging reactive oxygen species and by inhibiting DEP-induced
downregulation of Akt. Free Radical Biology & Medicine 39: 1549-1559.
Albert CM; Rosenthal L; Calkins H; Steinberg JS; Ruskin JN; Wang P; Muller JE; Mittleman MA. (2007). Driving and
Implantable Cardioverter-Defibrillator Shocks for Ventricular Arrhythmias: Results From the TOVA Study. Journal of
the American College of Cardiology 50: 2233-2240.
Albinet A; Leoz-Garziandia E; Budzinski H; Villenave E. (2007). Sampling precautions for the measurement of nitrated
polycyclic aromatic hydrocarbons in ambient air. Atmospheric Environment 41: 4988-4994.
Albrecht B. (1989). Aerosols, cloud microphysics and fractional cloudiness. Science 245: 1227-1230.
Alexeeva LB; Strachan WMJ; Shlychkova VV; Nazarova AA; Nikanorov AM; Korotova LG. (2001). Organochlorine pesticides
and trace metal monitoring of Russian rivers flowing to the Arctic Ocean. Marine Pollution Bulletin 43: 71-85.
Alexis A; Delao A; Garcia C; Nystrom M; Rosenkranz K. (2001). The 2001 California Almanac of Emissions and Air Quality.
California Air Resources Board, California Environmental Protection Agency, Sacramento, CA.
Alexis NE; Lay JC; Zeman K; Bennett WE; Peden DB; Soukup JM; Devlin RB; Becker S. (2006). Biological material on inhaled
coarse fraction particulate matter activates airway phagocytes in vivo in healthy volunteers. J Allergy Clin Immunol 117:
1396-1403.
Alfani A; Nicola FD; Maisto G; Prati MV. (2005). Long-term PAH accumulation after bud break in Quercus ilex L. leaves in a
polluted environment. Atmospheric Environment 39: 307-314.
Al-Horr R; Samanta G; Dasgupta PK. (2003). A Continuous Analyzer for Soluble Anionic Constituents and Ammonium in
Atmospheric Particulate Matter. ENVIRONMENTAL SCIENCE AND TECHNOLOGY-WASHINGTON DC- 37: 5711-
5720.
Alink GM; Sjogren M; Bos RP; Doekes G; Kromhout H; Scheepers PTJ. (1998). Effect of airborne particles from selected indoor
and outdoor environments on gap-junctional intercellular communication. Toxicology Letters 96: 209-213.
Allen GA; Harrison D; Koutrakis P. (2001). A New Method for Continuous Measurement of Sulfate in the Ambient Atmosphere.
Allen R; Liu LJ; Larson T; Sheppard L; Wallace L. (2003). Use of real-time light scattering data to estimate the contribution of
infiltrated and indoor-generated particles to indoor air. Environmental Science and Technology 37: 3484-3492.
Allen RW; Criqui MH; Diez Roux AV; Allison M; Shea S; Detrano R; Sheppard L; Wong N; Hinckley Stukovsky K; Kaufman
JD. (in press). Fine particulate air pollution, proximity to traffice, and aortic atherosclerosis: The Multi-Ethnic Study of
Atherosclerosis. Epidemiology.
Allen RW; Mar T; Koenig J; Liu LJ; Gould T; Simpson C; Larson T. (2008). Changes in lung function and airway inflammation
among asthmatic children residing in a woodsmoke-impacted urban area. Inhal Toxicol 20: 423-433.
Allison MA; Cheung P; Criqui MH; Langer RD; Wright CM. (2006). Mitral and aortic annular calcification are highly associated
with systemic calcified atherosclerosis. Circulation 113: 861-866.
Allison MA; Criqui MH; Wright CM. (2004). Patterns and risk factors for systemic calcified atherosclerosis. Arterioscler
Thromb Vase Biol 24: 331-336.
Almas AR; Bakken LR; Mulder J. (2004). Changes in tolerance of soil microbial communities in Zn and Cd contaminated soils.
Soil Biology and Biochemistry 36: 805-813.
Alves C; Pio C; Carvalho A; Santos C. (2006). Atmospheric carbonaceous aerosols over grasslands of central Europe and a
Boreal forest. Chemosphere 63: 153-164.
Amara N; Bachoual R; Desmard M; Golda S; Guichard C; Lanone S; Aubier M; Ogier-Denis E; Boczkowski J. (2007). Diesel
exhaust particles induce matrix metalloprotease-1 in human lung epithelial cells via a NADP(H) oxidase/NOX4 redox-
dependent mechanism. American journal of physiology Lung cellular and molecular physiology 293: L170-181.
Amato M; Montorsi P; Ravani A; Oldani E; Galli S; Ravagnani PM; Tremoli E; Baldassarre D. (2007). Carotid intima-media
thickness by B-mode ultrasound as surrogate of coronary atherosclerosis: correlation with quantitative coronary
angiography and coronary intravascular ultrasound findings. European heart journal 28: 2094-2101.
American Thoracic Society. (1985). Guidelines as to what constitutes an adverse respiratory health effect, with special reference
to epidemiologic studies of air pollution. Am Rev Respir Dis 131: 666-668.
American Thoracic Society. (2000). What constitutes an adverse health effect of air pollution? . Am J Respir Crit Care Med 161:
665-673.
Analitis A; Katsouyanni K; Dimakopoulou K; Samoli E; Nikoloulopoulos AK; Petasakis Y; Touloumi G; Schwartz J; Anderson
HR; Cambra K; Forastiere F; Zmirou D; Vonk JM; Clancy L; Kriz B; Bobvos J; Pekkanen J. (2006). Short-term effects of
ambient particles on cardiovascular and respiratory mortality. Epidemiology 17: 230-233.
December 2008
R-2
DRAFT-DO NOT QUOTE OR CITE
-------
Andersen ZJ; Loft S; Ketzel M; Stage M; Scheike T; Hermansen MN; Bisgaard H. (2008). Ambient air pollution triggers
wheezing symptoms in infants. Thorax 63: 710-716.
Andersen ZJ; Wahlin P; Raaschou-Nielsen O; Ketzel M; Scheike T; Loft S. (2007b). Size Distribution and Total Number
Concentration of Ultrafine and Accumulation Mode Particles and Hospital Admissions in Children and the Elderly in
Copenhagen, Denmark. Occupational and environmental medicine.
Andersen ZJ; Wahlin P; Raaschou-Nielsen O; Scheike T; Loft SCEH. (2007a). Ambient particle source apportionment and daily
hospital admissions among children and elderly in Copenhagen. Journal of Exposure Science and Environmental
Epidemiology 17: 625-636.
Anderson HR; Atkinson RW; Bremner SA; Marston L. (2003). Particulate air pollution and hospital admissions for
cardiorespiratory diseases: are the elderly at greater risk? Eur Respir J Suppl 40: 39s-46s.
Anderson HR; Bremner SA; Atkinson RW; Harrison RM; Walters S. (2001). Particulate matter and daily mortality and hospital
admissions in the west midlands conurbation of the United Kingdom: associations with fine and coarse particles, black
smoke and sulphate. Occupational and environmental medicine 58: 504-510.
Anderson ME; Bogdan GM. (2007). Environments, indoor air quality, and children. Pediatr ClinNorth Am 54: 295-307, viii.
Andracchio A; Cavicchi C; Tonelli D; Zappoli S. (2002). A new approach for the fractionation of water-soluble organic carbon in
atmospheric aerosols and cloud drops. Atmospheric Environment 36: 5097-5107.
Andreae MO; Gelencser A. (2006). Black carbon or brown carbon? The nature of light-absorbing carbonaceous aerosols. Atmos
Chem Phys 6: 3131-3148.
Andreae MO; Rosenfeld D; Artaxo P; Costa AA; Frank GP; Longo KM; Silva-Dias MAF. (2004). Atmospheric science: smoking
rain clouds over the Amazon. Science 303: 1337-1341.
Annesi-Maesano I; MoreauD; CaillaudD; LavaudF; Le Moullec Y; Taytard A; Pauli G; CharpinD. (2007). Residential
proximity fine particles related to allergic sensitisation and asthma in primary school children. Respir Med 101: 1721-
1729.
Anselm A; Heibel T; Gebhart J; Ferron GA. (1990). In vivo studies of growth factors of sodium chloride particles in the human
respiratory tract. J Aerosol Sci 21: S427-430.
Anselme F; Loriot S; Henry J-P; Dionnet F; Napoleoni J-G; Thnillez C; Morin J-P. (2007). Inhalation of diluted diesel engine
emission impacts heart rate variability and arrhythmia occurrence in a rat model of chronic ischemic heart failure.
Archives of toxicology 81: 299-307.
Anthony TR; Flynn MR. (2006). Computational fluid dynamics investigation of particle inhalability. J Aerosol Sci 37: 750-765.
Antonini JM; Roberts JR; Clarke RW; Yang HM; Barger MW; Ma JYC; WeissmanDN. (2001). Effect of Age on Respiratory
Defense Mechanisms* Pulmonary Bacterial Clearance in Fischer 344 Rats After Intratracheal Instillation of Listeria
monocytogenes (Vol. 120, pp. 240-249): Am Coll Chest Phys.
Antonini JM; Roberts JR; Jernigan MR; Yang HM; Ma JY; Clarke RW. (2002). Residual oil fly ash increases the susceptibility to
infection and severely damages the lungs after pulmonary challenge with a bacterial pathogen. Toxicol Sci 70: 110-119.
Antonini JM; Taylor MD; Leonard SS; Lawryk NJ; Shi X; Clarke RW; Roberts JR. (2004). Metal composition and solubility
determine lung toxicity induced by residual oil fly ash collected from different sites within a power plant. Molecular and
cellular biochemistry 255: 257-265.
Antoniou KM; Malagari K; Tzanakis N; Perisinakis K; Symvoulakis EK; Karkavitsas N; Siafakas NM; Bouros D. (2006).
Clearance of technetium-99m-DTPA and HRCT findings in the evaluation of patients with Idiopathic Pulmonary
Fibrosis. BMC Pulmonary Medicine 6: 4.
Appel K; Bhave PV; Gilliland AB; Sarwar G; Roselle SJ. (2008). Evaluation of the community multiscale air quality (CMAQ)
model version 4.5: Sensitivities impacting model performance; Part IIa€"particulate matter. Atmospheric Environment
42: 6057-6066.
Appel KW; Gilliland A; Eder B. (2005). An Operational Evaluation of the 2005 Release of Models-3 CMAQ Version 4.5.
National Oceanic and Atmospheric Administration-Air Resources.
Arad Y; Spadaro LA; Goodman K; Lledo-Perez A; Sherman S; Lerner G; Guerci AD. (1996). Predictive value of electron beam
computed tomography of the coronary arteries. 19-month follow-up of 1173 asymptomatic subjects. Circulation 93:
1951-1953.
Araujo JA; Barajas B; Kleinman M; Wang X; Bennett BJ; Gong KW; Navab M; Harkema J; Sioutas C; Lusis AJ; Nel AE.
(2008). Ambient particulate pollutants in the ultrafine range promote early atherosclerosis and systemic oxidative stress.
CircRes 102: 589-596.
Arbex MA; Martins LC; de Oliveira RC; Pereira LA; Arbex FF; Cancado JE; Saldiva PH; Braga AL. (2007). Air pollution from
biomass burning and asthma hospital admissions in a sugar cane plantation area in Brazil. Journal of epidemiology and
community health 61: 395-400.
Ardehali R; Nasir K; Kolandaivelu A; Budoff MJ; Blumenthal RS. (2007). Screening patients for subclinical atherosclerosis with
non-contrast cardiac CT. Atherosclerosis 192: 235-242.
Arena VC; Mazumdar S; Zborowski JV; Talbott EO; He S; Chuang YH; Schwerha JJ. (2006). A retrospective investigation of
PM10 in ambient air and cardiopulmonary hospital admissions in Allegheny County, Pennsylvania: 1995-2000. J Occup
Environ Med 48: 38-47.
Arhami M; Kuhn T; Fine PM; Delfino RJ; Sioutas C. (2006). Effects of Sampling Artifacts and Operating Parameters on the
Performance of a Semicontinuous Particulate Elemental Carbon/Organic Carbon Monitor. Environ Sci Technol 40: 945-
954.
December 2008
R-3
DRAFT-DO NOT QUOTE OR CITE
-------
Arlt VM; Glatt H; Gamboa da Costa G; Reynisson J; Takamura-Enya T; Phillips DH. (2007). Mutagenicity and DNA adduct
formation by the urban air pollutant 2-nitrobenzanthrone. Toxicol Sci 98: 445-457.
Armbrust DV; Retta A. (2002). Wind and sandblast damage to growing vegetation. Annals of Arid Zone 39: 273-284.
Arnold JR; Dennis RL; Tonnesen GS. (2003). Diagnostic evaluation of numerical air quality models with specialized ambient
observations: testing the Community Multiscale Air Quality modeling system (CMAQ) at selected SOS 95 ground sites.
Atmospheric Environment 37: 1185-1198.
Arnott WP; Moosmuller H; Sheridan PJ; Ogren JA; Raspet R; Slaton WV; Hand JL; Kreidenweis SM; Collett JL. (2003).
Photoacoustic and filter-based ambient aerosol light absorption measurements: Instrument comparisons and the role of
relative humidity. J Geophys Res 108: 4034.
Aroniadou-Anderjaska V; Fritsch B; Qashu F; Braga MF. (2008). Pathology and pathophysiology of the amygdala in
epileptogenesis and epilepsy. Epilepsy Res 78: 102-116.
Aschner M; Erikson KM; Dorman DC. (2005). Manganese Dosimetry: Species Differences and Implications for Neurotoxicity.
Critical reviews in toxicology 35: 1-32.
Asgharian B; Kelly JT; Tewksbury EW. (2003). Respiratory Deposition and Inhalability of Monodisperse Aerosols in Long-
Evans Rats (Vol. 71, pp. 104-111): Soc Toxicology.
Atiga WL; Calkins H; Lawrence JH; Tomaselli GF; Smith JM; Berger RD. (1998). Beat-to-Beat Repolarization Lability
Identifies Patients at Risk for Sudden Cardiac Death. Journal of Cardiovascular Electrophysiology 9: 899-908.
Atkinson R. (1994). Gas-phase tropospheric chemistry of organic compounds. Journal of Physical and Chemical Reference Data
2: 1-216.
Atkinson RW. (2004). Acute effects of air pollution on admissions: reanalysis of APHEA 2. American journal of respiratory and
critical care medicine 169: 1257-1258.
Atkinson RW; Anderson HR; Sunyer J; Ayres J; Baccini M; Vonk JM; Boumghar A; Forastiere F; Forsberg B; Touloumi G;
Schwartz J; Katsouyanni K. (2001). Acute effects of particulate air pollution on respiratory admissions: results from
APHEA 2 project. Air Pollution and Health: a European Approach. American journal of respiratory and critical care
medicine 164: 1860-1866.
Atkinson RW; Bremner SA; Anderson HR; Strachan DP; Bland JM; de Leon AP. (1999). Short-term associations between
emergency hospital admissions for respiratory and cardiovascular disease and outdoor air pollution in London. Archives
of environmental health 54: 398-411.
Auchincloss AH; Roux AV; Dvonch JT; Brown PL; Barr RG; Daviglus ML; Goff DC; Kaufman JD; O'Neill MS. (2008).
Associations between Recent Exposure to Ambient Fine Particulate Matter and Blood Pressure in the Multi-Ethnic Study
of Atherosclerosis (MESA). Environ Health Perspect 116: 486-491.
Aunan K; Fang J; Hu T; Seip HM; Vennemo H. (2006). Climate change and air quality - measures with co-benefits in China.
Environmental Science & Technology 40: 4822-4829.
Ausset P; Bannery F; Del Monte M; Lefevre RA. (1998). Recording of pre-industrial atmospheric environment by ancient crusts
on stone monuments. Atmospheric Environment 32: 2859-2863.
Auvray M; Bey I. (2005). Long-range transport to Europe: Seasonal variations and implications for the European ozone budget. J
Geophys Res 110.
Avila A; Rodrigo A. (2004). Trace metal fluxes in bulk deposition, throughfall and stemflow at two evergreen oakstands in NE
Spain subject to different exposure to the industrial environment. Atmospheric Environment 38: 171-180.
Avogbe PH; Ayi-Fanou L; Autrup H; Loft S; Fayomi B; Sanni A; Vinzents P; Moller P. (2005). Ultrafine particulate matter and
high-level benzene urban air pollution in relation to oxidative DNA damage. Carcinogenesis 26: 613-620.
Avol EL; Gauderman WJ; Tan SM; London SJ; Peters JM. (2001). Respiratory effects of relocating to areas of differing air
pollution levels. American journal of respiratory and critical care medicine 164: 2067-2072.
Ayers GP. (2004). Potential for simultaneous measurement of PM10, PM2. 5 and PM1 for air quality monitoring purposes using
a single TEOM. Atmospheric Environment 38: 3453-3458.
Ayres JG; Borm P; Cassee FR; Castranova V; Donaldson K; Ghio A; Harrison RM; Hider R; Kelly F; Kooter IM; Marano F;
Maynard RL; Mudway I; Nel A; Sioutas C; Smith S; Baeza-Squiban A; Cho A; Duggan S; Froines J. (2008). Evaluating
the Toxicity of Airborne Particulate Matter and Nanoparticles by Measuring Oxidative Stress Potential - A Workshop
Report and Consensus Statement. Inhalation Toxicology 20: 75 - 99.
Azimi S; Rocher V; Garnaud S; Varrault G; Thevenot DR. (2005). Decrease of atmospheric deposition of heavy metals in an
urban area from 1994 to 2002 (Paris, France). Chemosphere 61: 645-651.
Babin SM; Burkom HS; Holtry RS; Tabernero NR; Stokes LD; Davies-Cole JO; DeHaan K; Lee DH. (2007). Pediatric patient
asthma-related emergency department visits and admissions in Washington, DC, from 2001-2004, and associations with
air quality, socio-economic status and age group. Environ Health 6: 9.
Baccarelli A; Martinelli I; Zanobetti A; Grillo P; Hou LF; Bertazzi PA; Mannucci PM; Schwartz J. (2008). Exposure to
particulate air pollution and risk of deep vein thrombosis. Arch Intern Med 168: 920-927.
Baccarelli A; Zanobetti A; Martinelli I; Grillo P; Hou L; Giacomini S; Bonzini M; Lanzani G; Mannucci PM; Bertazzi PA;
Schwartz J. (2007b). Effects of exposure to air pollution on blood coagulation. J Thromb Haemost 5: 252-260.
Baccarelli A; Zanobetti A; Martinelli I; Grillo P; Hou L; Lanzini G; Mannucci PM; Bartazzi PA; Schwartz J. (2007a). Air
pollution, smoking, and plasma homocysteine. Environ Health Perspect 115: 176-181.
Bachelet D; Lenihan J; Neilson R; Drapek R; Kittel T. (2005). Simulating the response of natural ecosystems and their fire
regimes to climatic variability in Alaska. Can J For Res 35: 2244-2257.
December 2008
R-4
DRAFT-DO NOT QUOTE OR CITE
-------
Bachoual R; Boczkowski J; Goven D; Amara N; Tabet L; On D; Lecon-Malas V; Aubier M; Lanone S. (2007). Biological effects
of particles from the paris subway system. Chemical research in toxicology 20: 1426-1433.
Backe C; Cousins IT; Larsson P. (2004). PCB in soils and estimated soil-air exchange fluxes of selected PCB congeners in the
south of Sweden. Environmental Pollution 128: 59-72.
Backstrom M; Nilsson U; Hakansson K; Allard B; Karlsson S. (2003). Speciation of heavy metals in road runoff and roadside
total deposition. Water, Air, and Soil Pollution 147: 343-366.
Bae MS; Demeijian K; Schwab J; Weimer S; Hou J; Zhou X; Rhoads K; Orsini D. (2007a). Intercomparison of Real Time
Ammonium Measurements at Urban and Rural Locations in New York. Aerosol Science and Technology 41: 329-341.
Bae MS; Schauer JJ; Deminter JT; Turner JR. (2004). Hourly and Daily Patterns of Particle-Phase Organic and Elemental Carbon
Concentrations in the Urban Atmosphere. Journal of the Air & Waste Management Association 54: 823-833.
Bae MS; Schwab JJ; Zhang Q; Hogrefe O; Demeijian KL; Weimer S; Rhoads K; Orsini D; Venkatachari P; Hopke PK. (2007b).
Interference of organic signals in highly time resolved nitrate measurements by low mass resolution aerosol mass
spectrometry. J Geophys Res 112.
Bagate K; Meiring James J; Gerlofs-Nijland Miriam E; Vincent R; Cassee Flemming R; Borm Paul JA. (2004a). Vascular effects
of ambient particulate matter instillation in spontaneous hypertensive rats. Toxicology and applied pharmacology 197:
29-39.
Bagate K; Meiring JJ; Cassee FR; Borm PJA. (2004b). The effect of particulate matter on resistance and conductance vessels in
the rat. Inhalation toxicology 16: 431-436.
Bagate K; Meiring JJ; Gerlofs-Nijland ME; Cassee FR; Borm PJA. (2006b). Signal transduction pathways involved in particulate
matter induced relaxation in rat aorta—spontaneous hypertensive versus Wistar Kyoto rats. Toxicology in vitro 20: 52-62.
Bagate K; Meiring JJ; Gerlofs-Nijland ME; Cassee FR; Wiegand H; Osornio-Vargas A; Borm PJA. (2006a). Ambient particulate
matter affects cardiac recovery in a Langendorff ischemia model. Inhalation toxicology 18: 633-643.
Bailey R; Barrie LA; Halsall CJ; Fellin P; Muir DCG. (2000). Atmospheric organochlorine pesticides in the western Canadian
Arctic: evidence of trans-Pacific transport. Journal of Geophysical Research 105: 11805-11811.
Baker DG. (1998). Natural pathogens of laboratory mice, rats, and rabbits and their effects on research. Clin Microbiol Rev 11:
231-266.
Balashazy I; Hofmann W; Heistracher T. (1999). Computation of local enhancement factors for the quantification of particle
deposition patterns in airway bifurcations. J Aerosol Sci 30: 185-203.
Balashazy I; Hofmann W; Heistracher T. (2003). Local particle deposition patterns may play a key role in the development of
lung cancer. Journal of applied physiology 94: 1719-1725.
Balkanski Y; Schulz M; Claquin T; Moulin C; Ginoux P. (2003). Global emissions of mineral aerosol: formulation and validation
using satellite imagery. In Granier C, Artaxo P, Reeves CE (Eds.), Emission of Atmospheric Trace Compounds (pp. 239-
267). Amsterdam: Kluwer.
Ballester F; Iniguez C; Saez M; Perez-Hoyos S; Daponte A; Ordonez JM; Barcelo MA; Taracido M; Arribas F; Bellido J;
Cambra K; Canada A; Guillen JJ. (2003). Short-term relationship between air pollution and mortality in 13 Spanish cities.
Med Clin (Bare) 121: 684-689.
Ballester F; Rodriguez P; Iniguez C; Saez M; Daponte A; Galan I; Taracido M; Arribas F; Bellido J; Cirarda FB; Canada A;
Guillen JJ; Guillen-Grima F; Lopez E; Perez-Hoyos S; Lertxundi A; Toro S. (2006). Air pollution and cardiovascular
admissions association in Spain: results within the EMECAS project. Journal of epidemiology and community health 60:
328-336.
Ballester F; Saez M; Perez-Hoyos S; Iniguez C; Gandarillas A; Tobias A; Bellido J; Taracido M; Arribas F; Daponte A; Alonso
E; Canada A; Guillen-Grima F; Cirera L; Perez-Boillos MJ; Saurina C; Gomez F; Tenias JM. (2002). The EMECAM
project: a multicentre study on air pollution and mortality in Spain: combined results for particulates and for sulfur
dioxide. Occupational and environmental medicine 59: 300-308.
Bao L; Chen S; Wu L; Hei Tom K; Wu Y; Yu Z; Xu A. (2007). Mutagenicity of diesel exhaust particles mediated by cell-particle
interaction in mammalian cells. Toxicology 229: 91-100.
Baptista MS; Vasconcelos MTSD; Cabral JP; Freitas MC; Pacheco AMG. (2008). Copper, nickel and lead in lichen and tree bark
transplants over different periods of time. Environmental Pollution 151: 408-413.
Barn P; Larson T; Noullett M; Kennedy S; Copes R; Brauer M. (2008). Infiltration of forest fire and residential wood smoke: an
evaluation of air cleaner effectiveness. Journal of Exposure Science and Environmental Epidemiology 18: 503-511.
Barnett AG; Williams GM; Schwartz J; Best TL; Neller AH; Petroeschevsky AL; Simpson RW. (2006). The effects of air
pollution on hospitalizations for cardiovascular disease in elderly people in Australian and New Zealand cities. Environ
Health Perspect 114: 1018-1023.
Barnett AG; Williams GM; Schwartz J; Neller AH; Best TL; Petroeschevsky AL; Simpson RW. (2005). Air pollution and child
respiratory health: a case-crossover study in Australia and New Zealand. American journal of respiratory and critical care
medicine 171: 1272-1278.
Barraza-Villarreal A, Sunyer J, Hernandez-Cadena L, Escamilla-Nunez MC, Sienra-Monge JJ, Ramirez-Aguilar M,
et al. 2008. Air pollution, airway inflammation, and lung function in a cohort study of Mexico City
schoolchildren. Environ Health Perspect 116(6): 832-838.
Barregard L; Sallsten G; Andersson L; Almstrand AC; Gustafson P; Andersson M; Olin AC. (2008). Experimental exposure to
wood smoke: effects on airway inflammation and oxidative stress. Occupational and environmental medicine 65: 319-
324.
December 2008
R-5
DRAFT-DO NOT QUOTE OR CITE
-------
Barregard L; Sallsten G; Gustafson P; Andersson L; Johansson L; Basu S; Stigendal L. (2006). Experimental exposure to wood-
smoke particles in healthy humans: Effects on markers of inflammation, coagulation, and lipid peroxidation. Inhalation
Toxicology 18: 845-853.
Barreto RP; Albuquerque FC; Netto ADP. (2007). Optimization of an improved analytical method for the determination of 1-
nitropyrene in milligram diesel soot samples by high-performance liquid chromatography-mass spectrometry. Journal of
Chromatography A 1163: 219-227.
Barrett EG; Henson RD; Seilkop SK; McDonald JD; Reed MD. (2006). Effects of hardwood smoke exposure on allergic airway
inflammation in mice. Inhalation toxicology 18: 33-43.
Bartoli CR; Wellenius GA; Diaz EA; Lawrence J; Coull BA; Akiyama I; Lee LM; Okabe K; Verrier RL; Godleski JJ. (2008).
Mechanisms of Inhaled Fine Particulate Air Pollution-Induced Arterial Blood Pressure Changes. Endocrine 116.
Bartzokas A; Kassomenos P; Petrakis M; Celessides C. (2004). The effect of meteorological and pollution parameters on the
frequency of hospital admissions for cardiovascular and respiratory problems in Athens. Indoor and Built Environment
13: 271-275.
Bastain TM; Gilliland FD; Li YF; Saxon A; Diaz-Sanchez D. (2003). Intraindividual reproducibility of nasal allergic responses to
diesel exhaust particles indicates a susceptible phenotype. Clinical Immunology 109: 130-136.
Basu R; Feng WY; Ostro BD. (2008). Characterizing temperature and mortality in nine California counties. Epidemiology 19:
138-145.
Basu R; Woodruff TJ; Parker JD; Saulnier L; Schoendorf KC. (2004). Comparing exposure metrics in the relationship between
PM2.5 and birth weight in California. Journal of exposure analysis and environmental epidemiology 14: 391-396.
Batalha JR; Saldiva PH; Clarke RW; Coull BA; Stearns RC; Lawrence J; Murthy GG; Koutrakis P; Godleski JJ. (2002).
Concentrated ambient air particles induce vasoconstriction of small pulmonary arteries in rats. Environ Health Perspect
110: 1191-1197.
Bateson TF; Schwartz J. (2004). Who is sensitive to the effects of particulate air pollution on mortality? A case-crossover
analysis of effect modifiers. Epidemiology 15: 143-149.
Baumgardner RE; Isil SS; Bowser JJ; Fitzgerald KM. (1999). Measurements of rural sulfur dioxide and particle sulfate: Analysis
of CASTNet data, 1987 through 1996. Journal ofthe Air & Waste Management Association(1995) 49: 1266-1279.
Baxter LK; Clougherty JE; Laden F; Levy JI. (2007b). Predictors of concentrations of nitrogen dioxide, fine particulate matter,
and particle constituents inside of lower socioeconomic status urban homes. J ExpoSci EnvironEpidemiol 17: 433-444.
Baxter LK; Wright RJ; Levy JI; Clougherty JE; Paciorek CJ. (2007a). Predicting residential indoor concentrations of nitrogen
dioxide, fine particulate matter, and elemental carbon using questionnaire and geographic information system based data.
Atmospheric Environment 41: 6561-6571.
Bayer-Oglesby L; Grize L; Gassner M; Takken-Sahli K; Sennhauser FH; Neu U; Schindler C; Braun-Fahrlander C. (2005).
Decline of ambient air pollution levels and improved respiratory health in Swiss children. Environ Health Perspect 113:
1632-1637.
Baynard T; Garland RM; Ravishankara AR; Tolbert MA; Lovejoy ER. (2006). Key factors influencing the relative humidity
dependence of aerosol light scattering. Geophysical Research Letters 33 doi:10.1029/2005GL024898.
Bayram H; Ito K; Issa R; Ito M; Sukkar M; Chung KF. (2006). Regulation of human lung epithelial cell numbers by diesel
exhaust particles. Eur Respir J 27: 705-713.
BBC Research & Consulting. (2002). Phoenix Area Visibility Survey. Draft Report. Available at
http://www.azdeq.gov/environ/air/download/vis_021903f.pdf.
Beadsmoore C; Cheow HK; Szczepura K; Ruparelia P; Peters AM. (2007). Healthy passive cigarette smokers have increased
pulmonary alveolar permeability. Nuclear medicine communications 28: 75.
Becher R; Bucht A; Ovrevik J; Hongslo Jan K; Dahlman Hans J; Samuelsen Jan T; Schwarze Per E. (2007). Involvement of
NADPH Oxidase and iNOS in Rodent Pulmonary Cytokine Responses to Urban Air and Mineral Particles. Inhalation
toxicology 19: 645-655.
Becker S; Dailey LA; Soukup JM; Grambow SC; Devlin RB; Huang YC. (2005a). Seasonal variations in air pollution particle-
induced inflammatory mediator release and oxidative stress. Environmental health perspectives 113: 1032-1038.
Becker S; Fenton MJ; Soukup JM. (2002). Involvement of microbial components and toll-like receptors 2 and 4 in cytokine
responses to air pollution particles. Am J Respir Cell Mol Biol 27: 611-618.
Becker S; Mundandhara S; Devlin RB; Madden M. (2005b). Regulation of cytokine production in human alveolar macrophages
and airway epithelial cells in response to ambient air pollution particles: Further mechanistic studies. Toxicol Appl
Pharmacol 207: 269-275.
Becker T; Brandel M. (2007). Vegetation-environment relationships in a heavy metal-dry grassland complex. Folia Geobotanica
42: 11-28.
Beckett WS; Chalupa DF; Pauly-Brown A; Speers DM; Stewart JC; Frampton MW; Utell MJ; Huang LS; Cox C; Zareba W;
Oberdorster G. (2005). Comparing inhaled ultrafine versus fine zinc oxide particles in healthy adults - A human
inhalation study. Am J Respir Crit Care Med 171:1129-1135.
Beck-Speier I; Dayal N; Karg E; Maier KL; Schumann G; Schulz H; Semmler M; Takenaka S; Stettmaier K; Bors W; Ghio A;
Samet JM; Heyder J. (2005). Oxidative stress and lipid mediators induced in alveolar macrophages by ultrafine particles.
Free RadicBiolMed 38: 1080-1092.
Becquemin MH; Swift DL; Bouchikhi A; Roy M; Teillac A. (1991). Particle deposition and resistance in the noses of adults and
children. European Respiratory Journal 4: 694-702.
December 2008
R-6
DRAFT-DO NOT QUOTE OR CITE
-------
Becquemin MM; Bertholon IF; Bouchikhi A; Malarbet JL; Roy M. (1999). Oronasal Ventilation Partitioning in Adults and
Children: Effect on Aerosol Deposition in Airways. Radiation Protection Dosimetry 81: 221-228.
Bedeschi E; Campari C; Candela S; Collini G; Caranci N; Frasca G; Galassi C; Francesca G; Vigotti MA. (2007). Urban air
pollution and respiratory emergency visits at pediatric unit, Reggio Emilia, Italy. J Toxicol Environ Health A 70: 261-
265.
Beeby A; Richmond L. (2002). Evaluating Helix aspersa as a sentinel for mapping metal pollution. Ecological Indicators 1: 261-
270.
Beelen R; Hoek G; van den Brandt PA; Goldbohm RA; Fischer P; Schouten LJ; Jerrett M; Hughes E; Armstrong B; Brunekreef
B. (2008a). Long-term effects of traffic-related air pollution on mortality in a Dutch cohort (NLCS-AIR study). Environ
Health Perspect 116: 196-202.
Beelen R; Hoek G; van den Brandt PA; Goldbohm RA; Fischer P; Schouten LJ; Armstrong B; Brunekreef B. (2008b). Long-term
exposure to traffic-related air pollution and lung cancer risk. Epidemiology 19: 702-710.
Behndig AF; Mudway IS; Brown JL; Stenfors N; Helleday R; Duggan ST; Wilson SJ; Boman C; Cassee FR; Frew AJ; Kelly FJ;
Sandstrom T; Blomberg A. (2006). Airway antioxidant and inflammatory responses to diesel exhaust exposure in healthy
humans. European Respiratory Journal 27: 359-365.
Behrentz E; Fitz DR; Pankratz DV; Sabin LD; Colome SD; Fruin SA; Winer AM. (2004). Measuring self-pollution in school
buses using a tracer gas technique. Atmospheric Environment 38: 3735-3746.
Bein KJ; Zhao Y; Wexler AS; Johnston MV. (2005). Speciation of size-resolved individual ultrafine particles in Pittsburgh,
Pennsylvania. J Geophys Res 110.
Bell ML; Dominici F; Ebisu K; Zeger SL; Samet JM. (2007a). Spatial and Temporal Variation in PM2. 5 Chemical Composition
in the United States for Health Effects Studies. Environmental health perspectives 115: 989.
Bell ML; Ebisu K; Belanger K. (2007b). Ambient air pollution and low birth weight in Connecticut and Massachusetts. Environ
Health Perspect 115: 1118-1124.
Bell ML; Ebisu K; Peng RD; Walker J; Samet JM; Zeger SL; Dominic F. (2008a). Seasonal and regional short-term effects of
fine particles on hospital admissions in 202 U.S. counties, 1999-2005. Am J Epidemiol In Press.
Bell ML; Levy JK; Lin Z. (2008b). The effect of sandstorms and air pollution on cause-specific hospital admissions in Taipei,
Taiwan. Occupational and environmental medicine 65: 104-111.
Bell ML; Samet JM; Dominici F. (2004). Time-series studies of particulate matter. Annu Rev Public Health 25: 247-280.
Bellouin N; Boucher O; Haywood J; Reddy MS. (2005). Global emissions of aerosol direct radiative forcing from satellite
measurements. Nature 438: 1138-1141.
Benkovitz CM; Scholtz MT; Pacyna J; Tarrason L; Dignon J; Voldner EC; Spiro PA; Logan JA; Graedel TE. (1996). Global
gridded inventories of anthropogenic emissions of sulfur and nitrogen: North Atlantic Regional Experiment(NARE).
Journal of Geophysical Research 101: 29239-29253.
Bennett CM; McKendiy IG; Kelly S; Denike K; Koch T. (2006). Impact of the 1998 Gobi dust event on hospital admissions in
the Lower Fraser Valley, British Columbia. The Science of the total environment 366: 918-925.
Bennett CM; Simpson P; Raven J; Skoric B; Powell J; Wolfe R; Walters EH; Abramson MJ. (2007). Associations between
ambient PM2.5 concentrations and respiratory symptoms in Melbourne, 1998-2005. J Toxicol Environ Health A 70:
1613-1618.
Bennett DH; Koutrakis P. (2006). Determining the infiltration of outdoor particles in the indoor environment using a dynamic
model. J Aerosol Sci 37: 766-785.
Bennett WD; Ilowite JS. (1989). Dual pathway clearance of/sup 99m/Tc-DTPA from the bronchial mucosa. Am Rev Respir Dis
139.
Bennett WD; Zeman KL. (1998). Deposition of fine particles in children spontaneously breathing at rest. Inhalation Toxicology
10: 831-842.
Bennett WD; Zeman KL. (2004). Effect of body size on breathing pattern and fine-particle deposition in children. Journal of
applied physiology 97: 821-826.
Bennett WD; Zeman KL. (2005). Effect of Race on Fine Particle Deposition for Oral and Nasal Breathing. Inhalation Toxicology
17: 641-648.
Bennett WD; Zeman KL; Jarabek AM. (2008). Nasal contribution to breathing and fine particle deposition in children versus
adults. J Toxicol Environ Health A 71: 227-237.
Bennett WD; Zeman KL; Kim C. (1996). Variability of fine particle deposition in healthy adults: effect of age and gender.
American journal of respiratory and critical care medicine 153: 1641-1647.
Bennett WD; Zeman KL; Kim C; Mascarella J. (1997). Enhanced deposition of fine particles in COPD patients spontaneously
breathing at rest. Inhalation Toxicology 9: 1-14.
Berger A; Zareba W; Schneider A; Ruckerl R; Ibald-Mulli A; Cyrys J; Wichmann HE; Peters A. (2006). Runs of ventricular and
supraventricular tachycardia triggered by air pollution in patients with coronary heart disease. J Occup Environ Med 48:
1149-1158.
Berger RD; Kasper EK; Baughman KL; Marban E; Calkins H; Tomaselli GF. (1997). Beat-to-Beat QT Interval Variability :
Novel Evidence for Repolarization Lability in Ischemic and Nonischemic Dilated Cardiomyopathy. Circulation 96: 1557-
1565.
December 2008
R-7
DRAFT-DO NOT QUOTE OR CITE
-------
Bermudez E; Mangum IB; Asgharian B; Wong BA; Reverdy EE; Janszen DB; Hext PM; Warheit DB; Everitt JI. (2002). Long-
Term Pulmonary Responses of Three Laboratory Rodent Species to Subchronic Inhalation of Pigmentary Titanium
Dioxide Particles. Soc Toxicology 70: 86-97.
Bermudez E; Mangum JB; Wong BA; Asgharian B; Hext PM; Warheit DB; Everitt JI. (2004). Pulmonary Responses of Mice,
Rats, and Hamsters to Subchronic Inhalation of Ultrafine Titanium Dioxide Particles. Soc Toxicology 77: 347-357.
Beron K; Murdoch J; Thayer M. (2001). The Benefits of Visibility Improvement: New Evidence from the Los Angeles
Metropolitan Area. The Journal of Real Estate Finance and Economics 22: 319-337.
Bhalla DK; Mannix RC; Kleinman MT; Crocker TT. (1986). Relative permeability of nasal, tracheal, and bronchoalveolar
mucosa to macromolecules in rats exposed to ozone. Journal of Toxicology and Environmental Health 17: 269-283.
Bhattacharyya SN; Dubick MA; Yantis LD; Enriquez JI; Buchanan KC; Batra SK; Smiley RA. (2004). In vivo effect of wood
smoke on the expression of two mucin genes in rat airways. Inflammation 28: 67-76.
Bhavsar SP; Diamond ML; Evans LJ; Gandhi N; Nilsen J; Antunes P. (2004a). Development of a coupled metal speciation-fate
model for surface aquatic systems. Environmental Toxicology and Chemistry 23: 1376- 1385.
Bhavsar SP; Diamond ML; Gandhi N; Nilsen J. (2004b). Dynamic coupled metal transport-speciation model: application to
assess a zinc-contaminated lake. Environmental Toxicology and Chemistry 23: 2410-2420.
Bickerstaff K; Walker G. (2001). Public understandings of air pollution: the 'localisation'of environmental risk. Global Environ
Change 11: 133-145.
Bidleman TF; Jantunen LMM; Helm PA; Bronstrom-Lunden E; Juntto S. (2002). Chlordane isomers and enantiomers suggest
changing sources in Arctic air. Environmental Science and Technology 36: 539-544.
Bigazzi R; Bianchi S; Baldari D; Campese VM. (1998). Microalbuminuria predicts cardiovascular events and renal insufficiency
in patients with essential hypertension. JHypertens 16: 1325-1333.
Billings WD. (1978). Plants and the ecosystem (3rd ed.). Belmont, CA: Wadsworth Publishing Co., Inc.
Binkova B; Cerna M; Pastorkova A; Jeli'nek R; Benes I; Novak J; Sram RJ. (2003). Biological activities of organic compounds
adsorbed onto ambient air particles: comparison between the cities of Teplice and Prague during the summer and winter
seasons 2000-2001. Mutation Research-Fundamental and Molecular Mechanisms of Mutagenesis 525: 43-59.
Binkowski FS; Arunachalam S; Adelman Z; Pinto JP. (2007). Examining Photolysis Rates with a Prototype Online Photolysis
Module in CMAQ. Journal of Applied Meteorology and Climatology 46: 1252-1256.
Blanchard P; Brook JR; Brazal P. (2002). Chemical characterization of the organic fraction of atmospheric aerosol at two sites in
Ontario, Canada. Journal of Geophysical Research (Atmospheres) 107: D21.
Blanchet S; Ramgolam K; Baulig A; Marano F; Baeza-Squiban A. (2004). Fine particulate matter induces amphiregulin secretion
by bronchial epithelial cells. Am J Respir Cell Mol Biol 30: 421-427.
Blando JD; Turpin BJ. (2000). Secondary organic aerosol formation in cloud and fog droplets: a literature evaluation of
plausibility. Atmospheric Environment 34: 1623-1632.
Blank F; Rothen-Rutishauser B; Gehr P. (2007). Dendritic cells and macrophages form a transepithelial network against foreign
particulate antigens. American journal of respiratory cell and molecular biology 36: 669-677.
Bleck B; Tse Doris B; Jaspers I; de Lafaille Maria AC; Reibman J. (2006). Diesel exhaust particle-exposed human bronchial
epithelial cells induce dendritic cell maturation. Journal of immunology 176: 7431-7437.
Blomberg A; Sainsbury C; Rudell B; Frew AJ; Holgate ST; Sandstrom T; Kelly FJ. (1998). Nasal cavity lining fluid ascorbic
acid concentration increases in healthy human volunteers following short term exposure to diesel exhaust. Free Radical
Research 28: 59-67.
Bobak M. (2000). Outdoor air pollution, low birth weight, and prematurity. Environ Health Perspect 108: 173-176.
Bobak M; Leon DA. (1992). Air pollution and infant mortality in the Czech Republic, 1986-88. Lancet 340: 1010-1014.
Bobak M; Leon DA. (1999a). The effect of air pollution on infant mortality appears specific for respiratory causes in the
postneonatal period. Epidemiology 10: 666-670.
Bobak M; Leon DA. (1999b). Pregnancy outcomes and outdoor air pollution: an ecological study in districts of the Czech
Republic 1986-8. Occupational and environmental medicine 56: 539-543.
Bodolay E; Szekanecz Z; Devenyi K; Galuska L; Csipo I; Vegh J; Garai I; Szegedi G. (2005). Evaluation of interstitial lung
disease in mixed connective tissue disease (MCTD). Rheumatology 44: 656-661.
Boezen HM, van der Zee SC, Postma DS, Vonk JM, Gerritsen J, Hoek G, et al. 1999. Effects of ambient air
pollution on upper and lower respiratory symptoms and peak expiratory flow in children. Lancet 353(9156):
874-878.
Boezen HM; Vonk JM; van der Zee SC; Gerritsen J; Hoek G; Brunekreef B; Schouten JP; Postma DS. (2005). Susceptibility to
air pollution in elderly males and females. Eur Respir J 25: 1018-1024.
Bonazza A; Sabbioni C; Ghedini N. (2005). Quantitative data on carbon fractions in interpretation of black crusts and soiling on
European built heritage. Atmospheric Environment 39: 2607-2618.
Bond TC; Bergstrom RW. (2005). Light absorption by carbonaceous particles: an investigative review. Aerosol Science and
Technology 39: 1-41.
Bond TC; Streets DG; Yarber KF; Nelson SM; Woo J-H; Kilmont Z. (2004). A technology-based global inventory of black and
organic carbon emissions from combustion. Journal of Geophysical Research 109: 4203.
Bond TC; Sun H. (2005). Can reducing black carbon emissions counteract global warming? Environ Sci Technol 39: 5921-5926.
December 2008
R-8
DRAFT-DO NOT QUOTE OR CITE
-------
Bonner MR; Han D; Nie J; Rogerson P; Vena JE; Muti P; Trevisan M; Edge SB; Freudenheim JL. (2005). Breast cancer risk and
exposure in early life to polycyclic aromatic hydrocarbons using total suspended particulates as a proxy measure. Cancer
Epidemiol Biomarkers Prev 14: 53-60.
Borak J; Sirianni G; Cohen HJ; Chemerynski S; Wheeler R. (2003). Comparison of NIOSH 5040 Method versus Aethalometer™
to Monitor Diesel Particulate in School Buses and at Work Sites. AIHA Journal 64: 260-268.
Borrego C; Tchepel O; Costa AM; Martins H; Ferreira J; Miranda AI. (2006). Traffic-related particulate air pollution exposure in
urban areas. Atmospheric Environment 40: 7205-7214.
Bortnick SM; Coutant BW; Eberly SI. (2002). Using continuous PM2.5 monitoring data to report an air quality index. J Air
Waste Manag Assoc 52: 104-112.
Borys RD; Lowenthal DH; Cohn SA; Brown WOJ. (2003). Mountaintop and radar measurements of anthropogenic aerosol
effects on snow growth and snowfall rate. Geophysical Research Letters 30: 1538, doi: 1510.1029/2002GL016855.
Bosson J; Barath S; Pourazar J; Behndig AF; Sandstrom T; Blomberg A; Adelroth E. (2008). Diesel exhaust exposure enhances
the ozone-induced airway inflammation in healthy humans. Eur Respir J 31: 1234-1240.
Bosson J; Pourazar J; Forsberg B; Adelroth E; Sandstrom T; Blomberg A. (2007). Ozone enhances the airway inflammation
initiated by diesel exhaust. Respir Med 101:1140-1146.
Botter DA; Jorgensen B; Peres AA. (2002). A longitudinal study of mortality and air pollution for Sao Paulo, Brazil. Journal of
exposure analysis and environmental epidemiology 12: 335-343.
Boucher O; Reddy MS. (2008). Climate trade-off between black carbon and carbon dioxide emissions. Energy Policy 36: 193-
200.
Boucher U; Balabane M; Lamy I; Cambier P. (2005). Decomposition in soil microcosms of leaves of the metallophyte
Arabidopsis halleri: Effect of leaf-associated heavy metals on biodegradation. Environmental Pollution 135: 187-194.
Bourotte C; Curl-Amarante AP; Forti MC; Pereira LAA; Braga AL; Lotufo PA. (2007). Association between ionic composition
of fine and coarse aerosol soluble fraction and peak expiratory flow of asthmatic patients in Sao Paulo city (Brazil).
Atmospheric Environment 41: 2036-2048.
Bouthillier L; Vincent R; Goegan P; Adamson IY; Bjarnason S; Stewart M; Guenette J; Potvin M; Kumarathasan P. (1998).
Acute effects of inhaled urban particles and ozone: lung morphology, macrophage activity, and plasma endothelin-1. Am
J Pathol 153: 1873-1884.
Bracken MB; Triche EW; Belanger K; Saftlas A; Beckett WS; Leaderer BP. (2003). Asthma symptoms, severity, and drug
therapy: a prospective study of effects on 2205 pregnancies. Obstet Gynecol 102: 739-753.
Braga F; Mango JC; Souza JF; Ferrioli E; De Andrade J; Iazigi N. (1996). Age-related reduction in ? 99 ?Tc?m?-DTPA alveolar-
capillary clearance in normal humans. Nuclear medicine communications 17: 971-974.
Braude S; Nolop KB; Hughes JMB; Barnes PJ; Royston D. (1986). Comparison of lung vascular and epithelial permeability
indices in the adult respiratory distress syndrome. Am Rev Respir Dis 133: 1002-1005.
Brauer M; Gehring U; Brunekreef B; de Jongste J; Gerritsen J; Rovers M; Wichmann HE; Wijga A; Heinrich J. (2006). Traffic-
related air pollution and otitis media. EnvironHealth Perspect 114: 1414-1418.
Brauer M; Hoek G; Smit HA; de Jongste JC; Gerritsen J; Postma DS; Kerkhof M; Brunekreef B. (2007). Air pollution and
development of asthma, allergy and infections in a birth cohort. Eur Respir J 29: 879-888.
Brauer M; Hoek G; van Vliet P; Meliefste K; Fischer P; Gehring U; Heinrich J; Cyrys J; Bellander T; Lewne M; Brunekreef B.
(2003). Estimating long-term average particulate air pollution concentrations: application of traffic indicators and
geographic information systems. Epidemiology 14: 228-239.
Brauer M; Hoek G; Van Vliet P; Meliefste K; Fischer PH; Wijga A; Koopman LP; Neijens HJ; Gerritsen J; Kerkhof M. (2002).
Air pollution from traffic and the develo ment of respiratory infections and asthmatic an allergic symptoms in children.
American journal of respiratory and critical care medicine 166: 1092.
Brauer M; Lencar C; Tamburic L; Koehoorn M; Demers P; Karr C. (2008). A cohort study of traffic-related air pollution impacts
on birth outcomes. Environ Health Perspect 116: 680-686.
Brauner EV; Forchhammer L; Moller P; Barregard L; Gunnarsen L; Afshari A; Wahlin P; Glasius M; Dragsted LO; Basu S;
Raaschou-Nielsen O; Loft S. (2008). Indoor particles affect vascular function in the aged: an air filtration-based
intervention study. American journal of respiratory and critical care medicine 177: 419-425.
Brauner EV; Forchhammer L; Moller P; Simonsen J; Glasius M; Wahlin P; Raaschou-Nielsen O; Loft S. (2007). Exposure to
ultrafine particles from ambient air and oxidative stress-induced DNA damage. Environ Health Perspect 115: 1177-1182.
Briggs DJ; Collins S; Elliott P; Fischer P; Kingham S; Lebret E; Pryl K; Van Reeuwijk H; Smallbone K; Van der Veen A.
(1997). Mapping urban air pollution using GIS: A regression-based approach. International Journal of Geographical
Information Science 11: 699-718.
Briggs DJ; de Hoogh K; Morris C; Gulliver J. (2008). Effects of travel mode on exposures to particulate air pollution. Environ Int
34: 12-22.
Brimblecombe P; Grossi CM. (2005). Aesthetic thresholds and blackening of stone buildings. Science of the Total Environment
349: 175-189.
Brock CA; Hudson PK; Lovejoy ER; Sullivan A; Nowak JB; Huey LG; Cooper OR; Cziczo DJ; de Gouw J; Fehsenfeld FC.
(2004). Particle characteristics following cloud-modified transport from Asia to North America. J Geophys Res 109.
Brook JR; Burnett RT; Dann TF; Cakmak S; Goldberg MS; Fan X; Wheeler AJ. (2007). Further interpretation of the acute effect
of nitrogen dioxide observed in Canadian time-series studies. Journal of Exposure Science and Environmental
Epidemiology 17: S36.
December 2008
R-9
DRAFT-DO NOT QUOTE OR CITE
-------
Brook JR; Zhang L; Di-Giovanni F; Padro J. (1999). Description and evaluation of a model of deposition velocities for routine
estimates of air pollutant dry deposition over North America. Part I: model development. Atmospheric Environment 33:
5037-5051.
Brook RD; Brook JR; Urch B; Vincent R; Rajagopalan S; Silverman F. (2002). Inhalation of fine particulate air pollution and
ozone causes acute arterial vasoconstriction in healthy adults. Circulation 105: 1534-1536.
Brook RD; Franklin B; Cascio W; Hong YL; Howard G; Lipsett M; Luepker R; Mittleman M; Samet J; Smith SC; Tager I.
(2004). Air pollution and cardiovascular disease - A statement for healthcare professionals from the expert panel on
population and prevention science of the American Heart Association. Circulation 109: 2655-2671.
Brookshire D. (1979). Methods Development for Assessing Air Pollution Control Benefits. Volume 2: Experiments in Valuing
Nonmarket Goods. A Case Study of Alternative Benefit Measures of Air Pollution in the South Coast Air Basin of
Southern California. US Environmental Protection Agency, Washington, DC. EPA-600/6-79-0016.
Brown AS; Butterfield DM; Yardley RE; Quincey PG. (2006). Studies of the effect of humidity and other factors on some
different filter materials used for gravimetric measurements of ambient particulate matter. Atmospheric Environment 40:
4670-4678.
Brown DM; Donaldson K; Borm PJ; Schins RP; Dehnhardt M; Gilmour P; Jimenez LA; Stone V. (2004a). Calcium and ROS-
mediated activation of transcription factors and TNF-alpha cytokine gene expression in macrophages exposed to ultrafine
particles. American Journal of Physiology 286: L344-L353.
Brown DM; Donaldson K; Stone V. (2004b). Effects of PM10 in human peripheral blood monocytes and J774 macrophages.
Respiratory research 5.
Brown DM; Hutchison L; Donaldson K; Stone V. (2007). The effects of PM10 particles and oxidative stress on macrophages and
lung epithelial cells: modulating effects of calcium-signaling antagonists. American journal of physiology - Lung cellular
and molecular physiology 292: 1444-1451.
Brown JS. (2005). Particle inhalability at low wind speeds. Inhal Toxicol 17: 831-837.
Brown JS; Wilson WE; Grant LD. (2005). Dosimetric comparisons of particle deposition and retention in rats and humans.
Inhalation toxicology 17: 355-385.
Brown JS; Zeman KL; Bennett WD. (2001). Regional Deposition of Coarse Particles and Ventilation Distribution in Healthy
Subjects and Patients with Cystic Fibrosis. J Aerosol Med 14: 443-454.
Brown JS; Zeman KL; Bennett WD. (2002). Ultrafine Particle Deposition and Clearance in the Healthy and Obstructed Lung.
American journal of respiratory and critical care medicine 166: 1240.
Brunekreef B; Janssen NA; De Hartog JJ; Oldenwening M; Meliefste K; Hoek G; Lanki T; Timonen KL; Vallius M; Pekkanen J;
Van Grieken R. (2005). Personal, indoor, and outdoor exposures to PM2.5 and its components for groups of
cardiovascular patients in Amsterdam and Helsinki. ResRepHealth Efflnst: 1-70.
Bruno P; Caselli M; de Gennaro G; Tutino M. (2007). Determination of polycyclic aromatic hydrocarbons (PAHs) in particulate
matter collected with low volume samplers. Talanta 72: 1357-1361.
Buck S; Dodds W; Fisher J; Hart D; Parker A; Stevenson J; Watson V; Welch E. (2000). Nutrient criteria technical guidance
manual, rivers and streams: US Environmental Protection Agency, Office of Water. EPA-822-B-00-002,
Bunger J; Krahl J; Munack A; Ruschel Y; Schroder O; Emmert B; Westphal G; Muller M; Hallier E; Bruning T. (2007a). Strong
mutagenic effects of diesel engine emissions using vegetable oil as fuel. Arch Toxicol 81: 599-603.
Bunger J; Krahl J; Weigel A; Schroder O; Bruning T; Muller M; Hallier E; Westphal G. (2006). Influence of fuel properties,
nitrogen oxides, and exhaust treatment by an oxidation catalytic converter on the mutagenicity of diesel engine emissions.
Arch Toxicol 80: 540-546.
Bunger J; Schappler-Scheele B; Hilgers R; Hallier E. (2007b). A 5-year follow-up study on respiratory disorders and lung
function in workers exposed to organic dust from composting plants. Int Arch Occup Environ Health 80: 306-312.
Burch WM; Nemmar A; Hoet PHM; Thomeer M; Nemery B; Vanquickenborne B; Vanbilloen H; Mortelmans L; Hoylaerts MF;
Verbruggen A. (2002). Passage of Inhaled Particles Into the Blood Circulation in Humans* Response. Circulation 106:
el41.
Burchiel SW; Lauer FT; Dunaway SL; Zawadzki J; McDonald JD; Reed MD. (2005). Hardwood smoke alters murine splenic T
cell responses to mitogens following a 6-month whole body inhalation exposure. Toxicol Appl Pharmacol 202: 229-236.
Burchiel SW; Lauer FT; McDonald JD; Reed MD. (2004). Systemic immunotoxicity in AJ mice following 6-month whole body
inhalation exposure to diesel exhaust. Toxicology and applied pharmacology 196: 337-345.
Burke JM; Zufall MJ; Ozkaynak H. (2001). A population exposure model for particulate matter: case study results for PM (2.5)
in Philadelphia, PA. Journal of exposure analysis and environmental epidemiology 11: 470-489.
Burkhardt J; Kaiser H; Kappen L; Goldbach HE. (2002). The possible role of aerosols on stomatal conductivity for water vapour.
Basic and Applied Ecology 2: 351-364.
Burnett RT; Brook J; Dann T; Delocla C; Philips O; Cakmak S; Vincent R; Goldberg MS; Krewski D. (2000). Association
between particulate-and gas-phase components of urban air pollution and daily mortality in eight Canadian cities. Inhal
Toxicol 12: 15-39.
Burnett RT; Cakmak S; Brook JR; Krewski D. (1997). The role of particulate size and chemistry in the association between
summertime ambient air pollution and hospitalization for cardiorespiratory diseases. Environ Health Perspect 105: 614-
620.
Burnett RT; Dales R; Krewski D; Vincent R; Dann T; Brook JR. (1995). Associations between ambient particulate sulfate and
admissions to Ontario hospitals for cardiac and respiratory diseases. Am J Epidemiol 142: 15-22.
December 2008
R-10
DRAFT-DO NOT QUOTE OR CITE
-------
Burnett RT; Goldberg MS. (2003). Size-fractionated particulate mass and daily mortality in eight Canadian cities. In Boston M
(Ed.), Revised Analyses of Time-Series Studies of Air Pollution and Health. Special Report, (pp. 85-90): Health Effects
Institute.
Burnett RT; Smith-Doiron M; Stieb D; Cakmak S; Brook JR. (1999). Effects of particulate and gaseous air pollution on
cardiorespiratory hospitalizations. Archives of environmental health 54: 130-139.
Burnett RT; Stieb D; Brook JR; Cakmak S; Dales R; Raizenne M; Vincent R; Dann T. (2004). Associations between short-term
changes in nitrogen dioxide and mortality in Canadian cities. Archives of environmental health 59: 228-236.
Burt R; Wilson MA; Keck TJ; Dougherty BD; Strom DE; Lindahl JA. (2003b). Trace element speciation in selected smelter-
contaminated soils in Anaconda and Deer Lodge Valley, Montana, USA. Advances in Environmental Research 8: 51-67.
Burt R; Wilson MA; Mays MD; Lee CW. (2003a). Major and trace elements of selected pedons in the USA. J Environ Qual 32:
2109-2121.
Burtscher H. (2005). Physical characterization of particulate emissions from diesel engines: a review. J Aerosol Sci 36: 896-932.
Buser MD; Parnell Jr CB; Shaw BW; Lacey RE. (2007a). Particulate matter sampler errors due to the interaction of particle size
and sampler performance characteristics: Ambient PM2. 5 samplers. Transactions of the ASABE 50: 241-254.
Buser MD; Parnell Jr CB; Shaw BW; Lacey RE. (2007b). Particulate matter sampler errors due to the interaction of particle size
and sampler performance characteristics: Ambient PM 10 samplers. Transactions of the ASABE 50: 229-240.
Buser MD; Parnell Jr CB; Shaw BW; Lacey RE. (2007c). Particulate matter sampler errors due to the interaction of particle size
and sampler performance characteristics: Background and Theory. Transactions of the ASABE 50: 221-228.
Butler AJ; Andrew MS; Russell AG. (2003). Daily sampling Of PM2. 5 in Atlanta: Results of the first year of the Assessment of
Spatial Aerosol Composition. Journal of Geophysical Research-Part D-Atmospheres 108.
Buzorius G; Hameri K; Pekkanen J; Kulmala M. (1999). Spatial variation of aerosol number concentration in Helsinki city.
Atmospheric Environment 33: 553-565.
Bytnerowicz A; Miller PR; Olszyk DM. (1987a). Dry deposition of nitrate, ammonium and sulfate to a Ceanothus crassifolius
canopy and surrogate surfaces. Atmospheric Environment 21: 1749-1757.
Bytnerowicz A; Miller PR; Olszyk DM; Dawson PJ; Fox CA. (1987b). Gaseous and particulate air pollution in the San Gabriel
Mountains of southern California. Atmospheric Environment 21: 1805-1814.
Byun D; Schere KL. (2006). Review of the Governing Equations, Computational Algorithms, and Other Components of the
Models-3 Community Multiscale Air Quality (CMAQ) Modeling System. Applied Mechanics Reviews 59: 51.
Byun DW; Ching JKS. (1999). Science Algorithms of the EPA Models-3 Community Multiscale Air Quality (CMAQ) Modeling
System: US Environmental Protection Agency, Office of Research and Development.
Cabada JC; Rees S; Takahama S; Khlystov A; Pandis SN; Davidson CI; Robinson AL. (2004). Mass size distributions and size
resolved chemical composition of fine particulate matter at the Pittsburgh supersite. Atmospheric Environment 38: 3127-
3141.
Cadle SH; Mulawa PA; Hunsanger EC; Nelson K; Ragazzi RA; Barrett R; Gallagher GL; Lawson DR; Knapp KT; Snow R.
(1999). Composition of light-duty motor vehicle exhaust particulate matter in the Denver, Colorado Area. Environmental
science & technology 33: 2328-2339.
Cakmur RV; Miller RL; Perlwitz J; Koch D; Geogdzhayev IV; Ginoux P; Tegen I; Zender CS. (2006). Constraining the global
dust emission and load by minimizing the difference between the model and observations. Journal of Geophysical
Research 111 D06207, doi:06210.01029/02005JD005791.
Cakmur RV; Miller RL; Torres O. (2004). Incorporating the effect of small scale circulations upon dust emission in an AGCM.
Journal of Geophysical Research 109: D07201.
Calderon-Garciduenas L; Maronpot RR; Torres-Jardon R; Henriquez-Roldan C; Schoonhoven R; Acuna-Ayala H; Villarreal-
Calderon A; Nakamura J; Fernando R; Reed W; Azzarelli B; Swenberg JA. (2003). DNA damage in nasal and brain
tissues of canines exposed to air pollutants is associated with evidence of chronic brain inflammation and
neurodegeneration. Toxicologic pathology 31: 524-538.
Calderon-Garciduenas L; Mora-Tiscareno A; Ontiveros E; Gomez-Garza G; Barragan-Mejia G; Broadway J; Chapman S;
Valencia-Salazar G; Jewells V; Maronpot RR; Henriquez-Roldan C; Perez-Guille B; Torres-Jardon R; Herrit L; Brooks
D; Osnaya-Brizuela N; Monroy ME; Gonzalez-Maciel A; Reynoso-Robles R; Villarreal-Calderon R; Solt AC; Engle
RW. (2008). Air pollution, cognitive deficits and brain abnormalities: A pilot study with children and dogs. Brain Cogn.
Calderon-Garciduenas L; Vincent R; Mora-Tiscareno A; Franco-Lira M; Henriquez-Roldan C; Barragan-Mejia G; Garrido-
Garcia L; Camacho-Reyes L; Valencia-Salazar G; Paredes R; Romero L; Osnaya H; Villarreal-Calderon R; Torres-Jardon
R; Hazucha MJ; Reed W. (2007). Elevated plasma endothelin-1 and pulmonary arterial pressure in children exposed to
air pollution. Environ Health Perspect 115: 1248-1253.
Caligiuri G; Levy B; Pernow J; Thoren P; Hansson GK. (1999). Myocardial infarction mediated by endothelin receptor signaling
in hypercholesterolemic mice. Proc Natl Acad Sci U S A 96: 6920-6924.
Campbell A; Oldham M; Becaria A; Bondy SC; Meacher D; Sioutas C; Misra C; Mendez LB; Kleinman M. (2005). Particulate
matter in polluted air may increase biomarkers of inflammation in mouse brain. Neurotoxicology 26: 133-140.
Campen MJ; Babu NS; Helms GA; Pett S; Wernly J; Mehran R; McDonald JD. (2005). Nonparticulate components of diesel
exhaust promote constriction in coronary arteries from ApoE-/- mice. Toxicological sciences 88: 95-102.
Campen MJ; McDonald JD; Gigliotti AP; Seilkop SK; Reed MD; Benson JM. (2003). Cardiovascular effects of inhaled diesel
exhaust in spontaneously hypertensive rats. Cardiovascular toxicology 3: 353-361.
December 2008
R-11
DRAFT-DO NOT QUOTE OR CITE
-------
Campen MJ; McDonald ID; Reed MD; Seagrave J. (2006). Fresh gasoline emissions, not paved road dust, alter cardiac
repolarization in ApoE-/-mice. Cardiovascular toxicology 6: 199-210.
Camuffo D. (1995). Physical weathering of stones. The Science of the total environment 167: 1-14.
Caner B. (1994). Impaired lung epithelial permeability in diabetics detected by technetium-99m-DTPA aerosol scintigraphy (Vol.
35, pp. 204-206): Soc Nuclear Med.
Cao D; Tal TL; Graves LM; Gilmour I; Linak W; Reed W; Bromberg PA; Samet JM. (2007). Diesel exhaust particulate-induced
activation of Stat3 requires activities of EGFR and Src in airway epithelial cells. American journal of physiology Lung
cellular and molecular physiology 292: L422-L429.
Carlsten C; Kaufman JD; Trenga CA; Allen J; Peretz A; Sullivan JH. (2008). Thrombotic markers in metabolic syndrome
subjects exposed to diesel exhaust. Inhal Toxicol 20: 917-921.
Carlsten C; Kaufman Joel D; Peretz A; Trenga Carol A; Sheppard L; Sullivan Jeffrey H. (2007). Coagulation markers in healthy
human subjects exposed to diesel exhaust. Thrombosis research 120: 849-855.
Carvalho LM; Cajador I; Martins-Loujao MA. (2006). Arbuscular mycorrhizal fungi enhance root cadmium and copper
accumulation in the roots of the salt marsh plant Aster tripolium L. Plant and Soil 285: 161-169.
Cascio WE; Cozzi E; Hazarika S; Devlin RB; Henriksen RA; Lust RM; Van Scott MR; Wingard CJ. (2007). Cardiac and
vascular changes in mice after exposure to ultrafine particulate matter. Inhalation toxicology 19: 67-73.
Case MW; Williams R; Yeatts K; Chen F-L; Scott J; Svendsen E; Devlin RB. (2008). Evaluation of a direct personal coarse
particulate matter monitor. Atmospheric Environment 42: 4446-4452.
Cass GR. (1998). Organic molecular tracers for particulate air pollution sources. Trends in Analytical Chemistry 17: 356-366.
Cassee FR; Boere AJF; Fokkens PHB; Leseman DLAC; Sioutas C; Kooter IM; Dormans JAMA. (2005). Inhalation of
concentrated particulate matter produces pulmonary inflammation and systemic biological effects in compromised rats.
Journal of toxicology and environmental health Part A 68: 773-796.
CDC. (2004). The health consequences of smoking: a report of the Surgeon General. U.S. Department of Health and Human
Services, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health.
Available from: http://www.cdc.gov/tobacco/data_statistics/sgr/sgr_2004/chapters.htm, Atlanta, GA.
CDC. (2008a). Table 3-1: Current Asthma Population Estimates, in thousands by Age, United States: National Health Interview
Survey, 2006. Centers for Disease Control and Prevention. Environmental Hazards and Health Effects: Asthma Program,
from http://www.cdc.gov/asthma/nhis/06/table3-l.htm
CDC. (2008b). Table 4-1: Current Asthma Prevalence Percents by Age, United States: National Health Interview Survey, 2006.
Environmental Hazards and Health Effects: Asthma Program, from http://www.cdc.gov/asthma/nhis/06/table3-l.htm
Chadha TS. (1987). Oronasal distribution of ventilation during exercise in normal subjects and patients with asthma and rhinitis.
Chest 92: 1037-1041.
Chahine T; Baccarelli A; Litonjua A; Wright RO; Suh H; Gold DR; Sparrow D; Vokonas P; Schwartz J. (2007). Particulate air
pollution, oxidative stress genes, and heart rate variability in an elderly cohort. Environ Health Perspect 115: 1617-1622.
Chakrabarti B; Sioutas C; Fine PM; Delfino R. (2004). Performance evaluation of the active-flow personal DataRAM PM2.5
mass monitor (Thermo Anderson pDR-1200) designed for continuous personal exposure measurements. Atmospheric
Environment 38: 3329-3340.
Chambless LE; Heiss G; Folsom AR; Rosamond W; Szklo M; Sharrett AR; Clegg LX. (1997). Association of coronary heart
disease incidence with carotid arterial wall thickness and major risk factors: the Atherosclerosis Risk in Communities
(ARIC) Study, 1987-1993. Am J Epidemiol 146: 483-494.
Chameides WL; Yu H; Liu SC; Bergin M; Zhou X; Mearns L; Wang G; Kiang CS; Say lor RD; Luo C; Huang Y; Steiner A;
Giorgi F. (1999). Case study of the effects of atmospheric aerosols and regional haze on agriculture: an opportunity to
enhance crop yields in China through emission controls? Proceedings of the National Academy of Science 96: 13626-
13633.
Chan CC; Chuang KJ; Chen WJ; Chang WT; Lee CT; Peng CM. (2008). Increasing cardiopulmonary emergency visits by long-
range transported Asian dust storms in Taiwan. Environ Res 106: 393-400.
Chan CC; Chuang KJ; Chien LC; Chen WJ; Chang WT. (2006a). Urban air pollution and emergency admissions for
cerebrovascular diseases in Taipei, Taiwan. European heart journal 27: 1238-1244.
Chan CC; Chuang KJ; Shiao GM; Lin LY. (2004). Personal exposure to submicrometer particles and heart rate variability in
human subjects. Environ Health Perspect 112: 1063-1067.
Chan RC-F; Wang M; Li N; Yanagawa Y; Onoe K; Lee JJ; Nel AE. (2006b). Pro-oxidative diesel exhaust particle chemicals
inhibit LPS-induced dendritic cell responses involved in T-helper differentiation. Journal of allergy and clinical
immunology 118: 455-465.
Chang C-C; Chiu H-F; Wu Y-S; Li Y-C; Tsai M-L; Shen C-K; Yang C-Y. (2005a). The induction of vascular endothelial growth
factor by ultrafine carbon black contributes to the increase of alveolar-capillary permeability. Environmental health
perspectives 113: 454-460.
Chang C-C; Hwang J-S; Chan C-C; Cheng T-J. (2007a). Interaction effects of ultrafine carbon black with iron and nickel on heart
rate variability in spontaneously hypertensive rats. Environmental health perspectives 115: 1012-1017.
Chang C-C; Hwang J-S; Chan C-C; Wang P-Y; Cheng T-J. (2007b). Effects of concentrated ambient particles on heart rate,
blood pressure, and cardiac contractility in spontaneously hypertensive rats during a dust storm event. Inhalation
Toxicology 19: 973-978.
December 2008
R-12
DRAFT-DO NOT QUOTE OR CITE
-------
Chang C-C; Hwang J-S; Chan C-C; Wang P-Y; Hu T-H; Cheng T-J. (2005b). Effects of concentrated ambient particles on heart
rate variability in spontaneously hypertensive rats. Journal of occupational health 47: 471-480.
Chang CC; Hwang JS; Chan CC; Wang PY; Hu TH; Cheng TJ. (2004). Effects of concentrated ambient particles on heart rate,
blood pressure, and cardiac contractility in spontaneously hypertensive rats. Inhalation Toxicology 16: 421-429.
Chang CC; Tsai SS; Ho SC; Yang CY. (2005c). Air pollution and hospital admissions for cardiovascular disease in Taipei,
Taiwan. Environ Res 98: 114-119.
Chang CH; Meroney RN. (2003). Concentration and flow distributions in urban street canyons: wind tunnel and computational
data. Journal of Wind Engineering & Industrial Aerodynamics 91: 1141-1154.
Chang CT; Tsai CJ. (2003). A model for the relative humidity effect on the readings of the PM10 beta-gauge monitor. J Aerosol
Sci 34: 1685-1697.
Chang M-C; Geller MD; Sioutas C; Fokkens PHB; Cassee FR. (2002). Development and Evaluation of a Compact, Highly
Efficient Coarse Particle Concentrator for Toxicological Studies. Aerosol Science and Technology 36: 492 - 501.
Chardon B; Lefranc A; Granados D; Gremy I. (2007). Air pollution and doctors' house calls for respiratory diseases in the
Greater Paris area (2000-3). Occupational and environmental medicine 64: 320-324.
Charron A; Harrison RM; Moorcroft S; Booker J. (2004). Quantitative interpretation of divergence between PM10 and PM2.5
mass measurement by TEOM and gravimetric (Partisol) instruments. Atmospheric Environment 38: 415-423.
Chauhan V; Breznan D; Thomson E; Karthikeyan S; Vincent R. (2005). Effects of ambient air particles on the endothelin system
in human pulmonary epithelial cells (A549). Cell biology and toxicology 21: 191-205.
Che W; Zhang Z; Zhang H; Wu M; Liang Y; Liu F; Shu Y; Li N. (2007). Compositions and oxidative damage of condensate,
particulate and semivolatile organic compounds from gasoline exhausts. Environmental Toxicology and Pharmacology
24: 11-18.
Checkoway H; Levy D; Sheppard L; Kaufman J; Koenig J; Siscovick D. (2000). A case-crossover analysis of fine particulate
matter air pollution and out-of-hospital sudden cardiac arrest. Res Rep Health Eff Inst: 5-28; discussion 29-32.
Cheggour M; Chafik A; Fisher NS; Benbrahim S. (2005). Metal concentrations in sediments and clams in four Moroccan
estuaries. Marine Environmental Research 59: 119-137.
Chen CH; Xirasagar S; Lin HC. (2006a). Seasonality in adult asthma admissions, air pollutant levels, and climate: a population-
based study. J Asthma 43: 287-292.
Chen J; Tan M; Nemmar A; Song W; Dong M; Zhang G; Li Y. (2006b). Quantification of extrapulmonary translocation of
intratracheal-instilled particles in vivo in rats: Effect of lipopolysaccharide. Toxicology 222: 195-201.
Chen L; Yang W; JennisonBL; Goodrich A; Omaye ST. (2002). Air pollution and birth weight in northern Nevada, 1991-1999.
Inhal Toxicol 14: 141-157.
Chen LC; Hwang J-S. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice. IV.
Characterization of acute and chronic effects of ambient air fine particulate matter exposures on heart-rate variability.
Inhalation Toxicology 17: 209-216.
Chen LC; Nadziejko C. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice. V. CAPs
exacerbate aortic plaque development in hyperlipidemic mice. Inhalation Toxicology 17: 217-224.
Chen LH; Knutsen SF; Shavlik D; Beeson WL; Petersen F; Ghamsary M; Abbey D. (2005a). The association between fatal
coronary heart disease and ambient particulate air pollution: Are females at greater risk? Environ Health Perspect 113:
1723-1729.
Chen SJ; Lin WY; Huang KL; Lin CC; Hsieh LT; Kao MJ. (2004a). Characteristics of particles sampled in southern Taiwan
during the Asian dust storm periods in 2000 and 2001. Atmospheric Environment 38: 5925-5934.
Chen Y; Yang Q; Krewski D; Burnett RT; Shi Y; McGrail KM. (2005b). The effect of coarse ambient particulate matter on first,
second, and overall hospital admissions for respiratory disease among the elderly. Inhal Toxicol 17: 649-655.
Chen Y; Yang Q; Krewski D; Shi Y; Burnett RT; McGrail K. (2004b). Influence of relatively low level of particulate ar pollution
on hospitalization for COPD in elderly people. Inhal Toxicol 16: 21-25.
Cheng KH; Cheng YS; Yeh HC; Guilmette RA; Simpson SQ; Yang YH; Swift DL. (1996). In vivo measurements of nasal
airway dimensions and ultrafme aerosol deposition in the human nasal and oral airways. J Aerosol Sci 27: 785-801.
Cheng MF; Tsai SS; Wu TN; Chen PS; Yang CY. (2007). Air pollution and hospital admissions for pneumonia in a tropical city:
Kaohsiung, Taiwan. J Toxicol Environ Health A 70: 2021-2026.
Chernyak SM; Rice CP; McConnell LL. (1996). Evidence of currently used pesticides in air, ice, fog, seawater and surface
microlayer in the Bering and Chukchi Seas. Marine Pollution Bulletin 32: 410-419.
Chestnut LG; Dennis RL. (1997). Economic Benefits of Improvements in Visibility: Acid Rain Provisions of the 1990 Clean Air
Act Amendments. Journal of the Air & Waste Management Association 47: 395-402.
Chevalier P; Burri H; Adeleine P; Kirkorian G; Lopez M; Leizorovicz A; Andre-Fouet X; Chapon P; Rubel P; Touboul P. (2003).
QT dynamicity and sudden death after myocardial infarction: results of a long-term follow-up study. J Cardiovasc
Electrophysiol 14: 227-233.
Chiapello I; Moulin C; Prospero JM. (2005). Understanding the long-term variability of African dust transport across the Atlantic
as recorded in both Barbados surface concentrations and large-scale Total Ozone Mapping Spectrometer (TOMS) optical
thickness. J Geophys Res 110.
December 2008
R-13
DRAFT-DO NOT QUOTE OR CITE
-------
Chimonas MA; Gessner BD. (2007). Airborne particulate matter from primarily geologic, non-industrial sources at levels below
National Ambient Air Quality Standards is associated with outpatient visits for asthma and quick-relief medication
prescriptions among children less than 20 years old enrolled in Medicaid in Anchorage, Alaska. Environ Res 103: 397-
404.
Chinard FP. (1980). The alveolar-capillary barrier: some data and speculations. Microvasc Res 19: 1-17.
Chiou CT; Sheng G; Manes M. (2001). A partition-limited model for the plant uptake of organic contaminants from soil and
water. Environ Sci Technol 35: 1437-1444.
Cho AK; Sioutas C; Miguel AH; Kumagai Y; Schmitz DA; Singh M; Eiguren-Fernandez A; Froines JR. (2005a). Redox activity
of airborne particulate matter at different sites in the Los Angeles Basin. Environ Res 99: 40-47.
Cho H-Y; Jedlicka AE; Clarke R; Kleeberger SR. (2005b). Role of Toll-like receptor-4 in genetic susceptibility to lung injury
induced by residual oil fly ash. Physiological genomics 22: 108-117.
Cho H-Y; Reddy SP; Kleeberger SR. (2006). Nrf2 defends the lung from oxidative stress. Antioxidants & redox signaling 8: 76-
87.
Chock DP; Song Q; Hass H; Schell B; Ackermann I. (2003). Comment on "Control of fossil-fuel particulate black carbon and
organic matter, possibly the most effective method of slowing global warming" by Jacobson, M. Z. Journal of
Geophysical Research 108.
Choi JH; Xu QS; Park SY; Kim JH; Hwang SS; Lee KH; Lee HJ; Hong YC. (2007). Seasonal variation of effect of air pollution
on blood pressure. Journal of epidemiology and community health 61: 314-318.
Chow JC. (2007). The application of thermal methods for determining chemical composition of carbonaceous aerosols: A review.
Journal of Environmental Science and Health, Part A 42: 1521-1541.
Chow JC; Doraiswamy P; Watson JG; Chen LW; Ho SS; Sodeman DA. (2008). Advances in integrated and continuous
measurements for particle mass and chemical composition. J Air Waste Manag Assoc 58: 141-163.
Chow JC; Watson JG; Chen LW; Chang MC; Robinson NF; Trimble D; Kohl S. (2007). The IMPROVE_A temperature protocol
for thermal/optical carbon analysis: maintaining consistency with a long-term database. J Air Waste Manag Assoc 57:
1014-1023.
Chow JC; Watson JG; Chen LWA; Arnott WP; Moosmuller H; Fung K. (2004). Equivalence of Elemental Carbon by
Thermal/Optical Reflectance and Transmittance with Different Temperature Protocols. Environmental Science and
Technology 38: 4414-4422.
Chow JC; Watson JG; Louie PKK; Chen LWA; Sin D. (2005a). Comparison of PM2. 5 carbon measurement methods in Hong
Kong, China. Environmental Pollution 137: 334-344.
Chow JC; Watson JG; Lowenthal DH; Magliano KL. (2005b). Loss of PM2.5 nitrate from filter samples in central California. J
Air Waste ManagAssoc 55: 1158-1168.
Chow JC; Watson JG; Park K; Lowenthal DH; Robinson NF; Magliano KA. (2006). Comparison of Particle Light Scattering and
Fine Particulate Matter Mass in Central California. Journal of the Air & Waste Management Association 56: 398-410.
Christopher SA; Zhang J; Kaufman YJ; Remer L. (2006). Satellite-based assessment of the top of the atmosphere anthropogenic
aerosol radiative forcing over cloud-free oceans. Geophysical Research Letters 111, L15816,
doi:10.1029/2005GL025535.
ChuAJ. (2005). Tissue factor mediates inflammation. Archives of Biochemistry and Biophysics 440: 123-132.
Chuang KJ; Chan CC; Chen NT; Su TC; Lin LY. (2005a). Effects of particle size fractions on reducing heart rate variability in
cardiac and hypertensive patients. Environ Health Perspect 113: 1693-1697.
Chuang KJ; Chan CC; Shiao GM; Su TC. (2005b). Associations between submicrometer particles exposures and blood pressure
and heart rate in patients with lung function impairments. J Occup Environ Med 47: 1093-1098.
Chuang KJ; Chan CC; Su TC; Lee CT; Tang CS. (2007a). The effect of urban air pollution on inflammation, oxidative stress,
coagulation, and autonomic dysfunction in young adults. American journal of respiratory and critical care medicine 176:
370-376.
Chuang KJ; Chan CC; Su TC; Lin LY; Lee CT. (2007b). Associations between particulate sulfate and organic carbon exposures
and heart rate variability in patients with or at risk for cardiovascular diseases. J Occup Environ Med 49: 610-617.
Chuang KJ; Coull BA; Zanobetti A; Suh H; Schwartz J; Stone PH; Litonjua A; Speizer FE; Gold DR. (2008). Particulate Air
Pollution as a Risk Factor for ST-Segment Depression in Patients With Coronary Artery Disease. Circulation: 1314-1320.
Chung CE; Ramanathan V; Kim D; Podgorny IA. (2005). Global anthropogenic aerosol direct forcing derived from satellite and
ground-based observations. Journal of Geophysical Research 110, D24207, doi:10.1029/2005JD006356.
Chung SH; Seinfeld JH. (2002). Global distribution and climate forcing of carbonaceous aerosols. Journal of Geophysical
Research 107: 4407.
Churg A; Brauer M; del Carmen Avila-Casado M; Fortoul TI; Wright JL. (2003). Chronic exposure to high levels of particulate
air pollution and small airway remodeling. Environ Health Perspect 111: 714-718.
Churg A; Stevens B; Wright JL. (1998). Comparison of the uptake of fine and ultrafme Ti02 in a tracheal explant system.
American Journal of Physiology- Lung Cellular and Molecular Physiology 274: 81-86.
Churg A; Xie C; Wang X; Vincent R; Wang RD. (2005). Air pollution particles activate NF-kappaB on contact with airway
epithelial cell surfaces. Toxicol Appl Pharmacol 208: 37-45.
ChylekP; Wong J. (1995). Effect of absorbing aerosols on global radiation budget. Geophysical Research Letters 22: 929-931.
December 2008
R-14
DRAFT-DO NOT QUOTE OR CITE
-------
Ciencewicki J; Gowdy K; Krantz QT; Linak WP; Brighton L; Gilmour MI; Jaspers I. (2007). Diesel exhaust enhanced
susceptibility to influenza infection is associated with decreased surfactant protein expression. Inhalation toxicology 19:
1121-1133.
Claeys M; Vermeylen R; Pashynska V; Cafmeyer J; Guyon P; Andreae MO; Artaxo P; Maenhaut W; Graham B; Vas G; Wang
W. (2004). Formation of Secondary Organic Aerosols Through Photooxidation of Isoprene. Science 303: 1173-1176.
Clancy L; Goodman P; Sinclair H; Dockery DW. (2002). Effect of air-pollution control on death rates in Dublin, Ireland: an
intervention study. Lancet 360: 1210-1214.
Clarke JF; Edgerton ES; Martin BE. (1997). Dry deposition calculations for the Clean Air Status and Trends Network.
Atmospheric Environment 31: 3667-3678.
Clarke RW; Antonini JM; Hemenway DR; Frank R; Kleeberger SR; Jakab GJ. (2000). Inhaled particle-bound sulfate: effects on
pulmonary inflammatory responses and alveolar macrophage function. Inhal Toxicol 12: 169-186.
Clewell HJ; Crump KS. (2005). Quantitative Estimates of Risk for Noncancer Endpoints. An International Journal 25: 285-289.
Clifton VL; Giles WB; Smith R; Bisits AT; Hempenstall PA; Kessell CG; Gibson PG. (2001). Alterations of placental vascular
function in asthmatic pregnancies. American journal of respiratory and critical care medicine 164: 546-554.
CloughWS. (1975). The deposition of particles on moss and grass surfaces. Atmospheric Environment 9: 1113-1119.
Coleridge UM; Coleridge JCG. (1994). Pulmonary Reflexes: Neural Mechanisms of Pulmonary Defense. Annual Reviews in
Physiology 56: 69-91.
Collier TK. (2003). Forensic Ecotoxicology: Establishing Causality between Contaminants and Biological Effects in Field
Studies. Human and Ecological Risk Assessment 9: 259 - 266.
Collins RN; Merrington G; McLaughlin MJ; Morel JL. (2003). Organic ligand and pH effects on isotopically exchangeable
cadmium in polluted soils. Soil Science Society of America Journal 67: 112-121.
Conny JM; Klinedinst DB; Wight SA; Paulsen JL. (2003). Optimizing Thermal-Optical Methods for Measuring Atmospheric
Elemental (Black) Carbon: A Response Surface Study. Aerosol Science and Technology 37: 703-723.
Cooke WF; Liousse C; Cachier H; Feichter J. (1999). Construction of a 1 ° x 1 ° fossil fuel emission data set for carbonaceous
aerosol and implementation and radiative impact in the ECHAM 4 model. Journal of Geophysical Research 104: 22137-
22162.
Corburn J. (2007). Urban land use, air toxics and public health: Assessing hazardous exposures at the neighborhood scale.
Environmental Impact Assessment Review 27: 145-160.
Corey LM; Baker C; Luchtel Daniel L. (2006). Heart-rate variability in the apolipoprotein E knockout transgenic mouse
following exposure to Seattle particulate matter. Journal of toxicology and environmental health Part A 69: 953-965.
Courtois A; Andujar P; Ladeiro Y; Baudrimont I; Delannoy E; Leblais V; Begueret H; Galland MAB; Brochard P; Marano F.
(2008). Impairment of NO-Dependent Relaxation in Intralobar Pulmonary Arteries: Comparison of Urban Particulate
Matter and Manufactured Nanoparticles. Environmental health perspectives 116: 1294.
Couto JA; Fernandez JA; Aboal JR; Carballeira A. (2004). Active biomonitoring of element uptake with terrestrial mosses: a
comparison of bulk and dry deposition. Science of the Total Environment 324: 211-222.
Cozzi E; Hazarika S; Stallings Howard W; Cascio Wayne E; Devlin Robert B; Lust Robert M; Wingard Christopher J; Van Scott
Michael R. (2006). Ultrafme particulate matter exposure augments ischemia-reperfusion injury in mice. American journal
of physiology Heart and circulatory physiology 291: H894-903.
Craven TE; Ryu JE; Espeland MA; Kahl FR; McKinney WM; Toole JF; McMahan MR; Thompson CJ; Heiss G; Crouse JR, 3rd.
(1990). Evaluation of the associations between carotid artery atherosclerosis and coronary artery stenosis. A case-control
study. Circulation 82: 1230-1242.
Crawford DW; Lipsen MS; Purdie DA; Lohan MC; Statham PJ; Whitney FA; Putland JN; Johnson WK; Sutherland N; Peterson
TD; Harrison PJ; Wong CS. (2003). Influence of zinc and iron enrichments on phytoplankton growth in the northeastern
subarctic Pacific. Limnology and Oceanography 48: 1583-1600.
Creasia DA; Poggenburg Jr JK; Nettesheim P. (1976). Elution of benzo [alpha] pyrene from carbon particles in the respiratory
tract of mice. J Toxicol Environ Health 1: 967-975.
Creighton PJ; Lioy PJ; Haynie FH; Lemmons TJ; Miller JL; Gerhart J. (1990). Soiling by atmospheric aerosols in an urban
industrial area. Journal of the Air and Waste Management Association 40: 1285-1289.
Crepineau C; Rychen G; Feidt C; Le Roux Y; Lichtfouse E; Laurent F. (2003). Contamination of pastures by polycyclic aromatic
hydrocarbons (PAHs) in the vicinity of a highway. Journal of Agricultural and Food Chemistry 51: 4841-4845.
Crimmins BS; Baker JE. (2006). Improved GC/MS methods for measuring hourly PAH and nitro-PAH concentrations in urban
particulate matter. Atmospheric Environment 40: 6764-6779.
Crist KC; Liu B; Kim M; Deshpande SR; John K. (2008). Characterization of fine particulate matter in Ohio: Indoor, outdoor,
and personal exposures. EnvironRes 106: 62-71.
Croteau MN; Luoma SN; Stewart AR. (2005). Trophic transfer of metals along freshwater food webs: Evidence of cadmium
biomagnification in nature. Limnology and Oceanography 50: 1511-1519.
Crump KS; Hoel DG; Langley CH; Peto R. (1976). Fundamental carcinogenic processes and their implications for low dose risk
assessment. Cancer Research 36: 2973-2979.
Cruts B; van Etten L; Tornqvist H; Blomberg A; Sandstrom T; Mills NL; Borm PJ. (2008). Exposure to diesel exhaust induces
changes in EEG in human volunteers. Part Fibre Toxicol 5: 4.
Cuny D; Denayer FO; De Foucault B; Schumacker R; Colein P; Van Haluwyn C. (2004). Patterns of metal soil contamination
and changes in terrestrial cryptogamic communities. Environmental Pollution 129: 289-297.
December 2008
R-15
DRAFT-DO NOT QUOTE OR CITE
-------
Cyrys J; Heinrich J; Hoek G; Meliefste K; Lewne M; Gehring U; Bellander T; Fischer P; van Vliet P; Brauer M. (2003).
Comparison between different traffic-related particle indicators: Elemental carbon (EC), PM 2.5 mass, and absorbance.
Journal of exposure analysis and environmental epidemiology 13: 134-143.
Dai YT; Juang YJ; Wu Y; Breysse PN; Hsu DJ. (2006). In vivo measurements of inhalability of ultralarge aerosol particles in
calm air by humans. J Aerosol Sci 37: 967-973.
Dales R. (2004). Ambient carbon monoxide may influence heart rate variability in subjects with coronary artery disease. J Occup
Environ Med 46: 1217-1221.
Dales R; Burnett RT; Smith-Doiron M; Adm D; Stieb DM; Brook JR. (2004). Air pollution and sudden infant death syndrome.
Pediatrics 113: 628-631.
Dales R; Liu L; Szyszkowicz M; Dalipaj M; Willey J; Kulka R; Ruddy TD. (2007). Particulate air pollution and vascular
reactivity: the bus stop study. International Archives of Occupational and Environmental Health 81: 159-164.
Dales R; Wheeler A; Mahmud M; Frescura AM; Smith-Doiron M; Nethery E; Liu L. (2008). The Influence of Living Near
Roadways on Spirometry and Exhaled Nitric Oxide in Elementary Schoolchildren. Environmental health perspectives
116: 1423.
Dales RE; Cakmak S; Doiron MS. (2006). Gaseous air pollutants and hospitalization for respiratory disease in the neonatal
period. Environ Health Perspect 114: 1751-1754.
Daniels MJ; Dominici F; Zeger SL; Samet JM. (2004). The National Morbidity, Mortality, and Air Pollution Study. Part III:
PM10 concentration-response curves and thresholds for the 20 largest US cities. Res Rep Health Eff Inst: 1-21;
discussion 23-30.
Danielsen PH; Brauner EV; Barregard L; Sallsten G; Wallin M; Olinski R; Rozalski R; Moller P; Loft S. (2008b). Oxidatively
damaged DNA and its repair after experimental exposure to wood smoke in healthy humans. Mutat Res 642: 37-42.
Danielsen PH; Risom L; Wallin H; Autrup H; Vogel U; Loft S; Moller P. (2008a). DNA damage in rats after a single oral
exposure to diesel exhaust particles. Mutation research 637: 49-55.
Dasch JM. (1987). Measurement of dry deposition to surfaces in deciduous and pine canopies. Environmental Pollution 44: 261-
277.
Dasgupta PK; Idowu AD; Li J. (2007). WO/2007/081635. Method and Apparatus for Analyzing Arsenic Concentrations Using
Gas Phase Ozone Chemiluminescence
Davidson CI; Wu Y-L. (1990). Dry deposition of particles and vapors. In Lindberg SE, Page AL, Norton SA (Eds.), Acidic
precipitation: vol. 3, sources, deposition, and canopy interactions (pp. 103-216). New York: Springer-Verlag, Inc.
Davis AP; Shokouhian M; Ni S. (2001). Loading estimates of lead, copper, cadmium, and zinc in urban runoff from specific
sources. Chemosphere 44: 997-1009.
Davis MRH; Zhao FJ; McGrath SP. (2004). Pollution-induced community tolerance of soil microbes in response to a zinc
gradient. Environmental Toxicology and Chemistry 23: 2665-2672.
Davison PS; Roberts DL; Arnold RT; Colvile RN. (2004). Estimating the direct radiative forcing due to haze from the 1997
forest fires in Indonesia. Journal of Geophysical Research 109.
Day R. (2007). Place and the experience of air quality. Health and Place 13: 249-260.
De Bruin ML; van Hemel NM; Leufkens HG; Hoes AW. (2005). Hospital discharge diagnoses of ventricular arrhythmias and
cardiac arrest were useful for epidemiologic research. Journal of clinical epidemiology 58: 1325-1329.
de Haar C; Hassing I; Bol M; Bleumink R; Pieters R. (2005). Ultrafine carbon black particles cause early airway inflammation
and have adjuvant activity in a mouse allergic airway disease model. Toxicological sciences 87: 409-418.
de Haar C; Hassing I; Bol M; Bleumink R; Pieters R. (2006). Ultrafine but not fine particulate matter causes airway inflammation
and allergic airway sensitization to co-administered antigen in mice. Clinical and experimental allergy - journal of the
British Society for Allergy and Clinical Immunology 36: 1469-1479.
de Hartog JJ; Hoek G; Peters A; Timonen KL; Ibald-Mulli A; Brunekreef B; Heinrich J; Tiittanen P; van Wijnen JH; Kreyling
W; Kulmala M; Pekkanen J. (2003). Effects of fine and ultrafine particles on cardiorespiratory symptoms in elderly
subjects with coronary heart disease: the ULTRA study. Am J Epidemiol 157: 613-623.
de Kok TM; Hogervorst JG; Briede JJ; van Herwijnen MH; Maas LM; Moonen EJ; Driece HA; Kleinjans JC. (2005).
Genotoxicity and physicochemical characteristics of traffic-related ambient particulate matter. Environ Mol Mutagen 46:
71-80.
De Leon SF; Thurston GD; Ito K. (2003). Contribution of respiratory disease to nonrespiratory mortality associations with air
pollution. American journal of respiratory and critical care medicine 167: 1117-1123.
De Nicola F; Maisto G; Prati MV; Alfani A. (2005). Temporal variations in PAH concentrations in Quercus ilex L. (holm oak)
leaves in an urban area. Chemosphere 61: 432-440.
DeBell L (2006). Spatial and seasonal patterns and temporal variability of haze and its constituents in the United States: Report
IV. Journal. Retrieved from http://vista.cira.colostate.edu/improve/Publications/Reports/2006/2006.htm
Decesari S; Facchini MC; Fuzzi S; Tagliavini E. (2000). Characterization of water-soluble organic compounds in atmospheric
aerosol: A new approach. Journal of Geophysical Research 105: 1481-1489.
Deckert T; Yokoyama H; Mathiesen E; Ronn B; Jensen T; Feldt-Rasmussen B; Borch-Johnsen K; Jensen JS. (1996). Cohort
study of predictive value of urinary albumin excretion for atherosclerotic vascular disease in patients with insulin
dependent diabetes. BMJ 312: 871-874.
Delfino RJ; Gong H, Jr.; Linn WS; Pellizzari ED; Hu Y. (2003a). Asthma symptoms in Hispanic children and daily ambient
exposures to toxic and criteria air pollutants. Environ Health Perspect 111: 647-656.
December 2008
R-16
DRAFT-DO NOT QUOTE OR CITE
-------
Delfino RJ; Gong H; Linn WS; Hu Y; Pellizzari ED. (2003b). Respiratory symptoms and peak expiratory flow in children with
asthma in relation to volatile organic compounds in exhaled breath and ambient air. Journal of exposure analysis and
environmental epidemiology 13: 348-363.
Delfino RJ; Quintana PJ; Floro J; Gastanaga VM; Samimi BS; Kleinman MT; Liu LJ; Bufalino C; Wu CF; McLaren CE. (2004).
Association of FEV1 in asthmatic children with personal and microenvironmental exposure to airborne particulate matter.
Environ Health Perspect 112: 932-941.
Delfino RJ; Staimer N; Gillen D; Tjoa T; Sioutas C; Fung K; George SC; Kleinman MT. (2006). Personal and ambient air
pollution is associated with increased exhaled nitric oxide in children with asthma. Environ Health Perspect 114: 1736-
1743.
Delfino RJ; Staimer N; Tjoa T; Polidori A; Arhami M; Gillen DL; Kleinman MT; Vaziri ND; Longhurst J; Zaldivar F; Sioutas C.
(2008). Circulating biomarkers of inflammation, antioxidant activity, and platelet activation are associated with primary
combustion aerosols in subjects with coronary artery disease. Environ Health Perspect 116: 898-906.
Delfino RJ; Zeiger RS; Seltzer JM; Street DH; McLaren CE. (2002). Association of asthma symptoms with peak particulate air
pollution and effect modification by anti-inflammatory medication use. Environ Health Perspect 110: A607-617.
DeLorenzo AJD. (1970). The olfactory neuron and the blood-brain barrier. Paper presented at the Taste and Smell in Vertebrates:
A Ciba Foundation Symposium.
DeMarini DM; Brooks LR; Warren SH; Kobayashi T; Gilmour MI; Singh P. (2004). Bioassay-directed fractionation and
salmonella mutagenicity of automobile and forklift diesel exhaust particles. Environ Health Perspect 112: 814-819.
DeMeo DL; Zanobetti A; Litonjua AA; Coull BA; Schwartz J; Gold DR. (2004). Ambient air pollution and oxygen saturation.
American journal of respiratory and critical care medicine 170: 383-387.
Demeijian KL; Mohnen VA. (2008). Synopsis of the temporal variation of particulate matter composition and size. J Air Waste
Manag Assoc 58: 216-233.
Denman KL; Brasseur G; Chidthaisong A; Ciais P; Cox PM; Dickinson RE; Hauglustaine D; Heinze C; Holland E; Jacob D;
Lohmann U; Ramachandran S; da Silva Dias PL; Wofsy SC; Zhang X. (2007). Couplings Between Changes in the
Climate System and Biogeochemistry. In Solomon S, Qin D, Manning M, Chen Z, Marquis M, Averyt KB, M.Tignor,
Miller HL (Eds.), Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth
Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge and New York: Cambridge University
Press.
Dentener F, et al. (2006). The global atmospheric environment for the next generation. Environmental Science and Technology
40: 3586-3594.
Dermentzoglou M; Manoli E; Voutsa D; Samara C. (2003). Sources and patterns of polycyclic aromatic hydrocarbons and heavy
metals in fine indoor particulate matter of Greek houses. Fresenius Environmental Bulletin 12: 1511-1519.
Derome J; Lindroos A-J. (1998). Effects of heavy metal contamination on macronutrient availability and acidification parameters
in forest soil in the vicinity of the Haijavalta Cu-Ni smelter, SW Finland. Environmental Pollution 99: 225-232.
Derwent RG; Collins WJ; Johnson CE; Stevenson DS. (2001). Transient Behaviour of Tropospheric Ozone Precursors in a
Global 3-D CTM and Their Indirect Greenhouse Effects. Climatic Change 49: 463-487.
Desqueyroux H; Pujet JC; Prosper M; Le Moullec Y; Momas I. (2002b). Effects of air pollution on adults with chronic
obstructive pulmonary disease. Archives of environmental health 57: 554-560.
Desqueyroux H; Pujet JC; Prosper M; Squinazi F; Momas I. (2002a). Short-Term Effects of Low-Level Air Pollution on
Respiratory Health of Adults Suffering from Moderate to Severe Asthma. Environmental Research 89: 29-37.
Deurloo DT; van Esch BC; Hofstra CL; Nijkamp FP; van Oosterhout AJ. (2001). CTLA4-IgG reverses asthma manifestations in
a mild but not in a more "severe" ongoing murine model. Am J Respir Cell Mol Biol 25: 751-760.
Devlin RB; Ghio AJ; Kehrl H; Sanders G; Cascio W. (2003). Elderly humans exposed to concentrated air pollution particles have
decreased heart rate variability. European Respiratory Journal 21: 76S-80S.
Di Toro DM; McGrath JH; Berry WJ; Paquin PR; Mathew R; Wu KB; Santore RC. (2005). Predicting sediment metal toxicity
using a sediment biotic ligand model: Methodology and initial application. Environmental Toxicology and Chemistry 24:
2410-2427.
Diapouli E; Chaloulakou A; Spyrellis N. (2007). Levels of ultrafine particles in different microenvironments—implications to
children exposure. Sci Total Environ 388: 128-136.
Diaz-Sanchez D; Penichet Garcia M; Wang M; Jyrala M; Saxon A. (1999). Nasal challenge with diesel exhaust particles can
induce sensitization to a neoallergen in the human mucosa. Journal of allergy and clinical immunology 104: 1183-1188.
Diaz-Sanchez D; Tsien A; Fleming J; Saxon A. (1997). Combined diesel exhaust particulate and ragweed allergen challenge
markedly enhances human in vivo nasal ragweed-specific IgE and skews cytokine production to a T helper cell 2-type
pattern. J Immunol 158: 2406-2413.
Dick CA; Singh P; Daniels M; Evansky P; Becker S; Gilmour MI. (2003a). Murine pulmonary inflammatory responses following
instillation of size-fractionated ambient particulate matter. J Toxicol Environ Health A 66: 2193-2207.
Dick CAJ; Brown DM; Donaldson K; Stone V. (2003b). The role of free radicals in the toxic and inflammatory effects of four
different ultrafine particle types. Inhalation toxicology 15: 39-52.
Diez Roux AV; Auchincloss AH; Astor B; Barr RG; Cushman M; Dvonch T; Jacobs DR, Jr.; Kaufman J; Lin X; Samson P.
(2006). Recent exposure to particulate matter and C-reactive protein concentration in the multi-ethnic study of
atherosclerosis. Am J Epidemiol 164: 437-448.
December 2008
R-17
DRAFT-DO NOT QUOTE OR CITE
-------
Diez Roux AV; Auchincloss AH; Franklin TG; Raghunathan T; Barr RG; Kaufman J; Astor B; Keeler J. (2008). Long-term
exposure to ambient particulate matter and prevalence of subclinical atherosclerosis in the Multi-Ethnic Study of
Atherosclerosis. Am J Epidemiol 167: 667-675.
Dinneen SF; Gerstein HC. (1997). The association of microalbuminuria and mortality in non-insulin-dependent diabetes mellitus.
A systematic overview of the literature. Arch Intern Med 157: 1413-1418.
DTppoliti D; Forastiere F; Ancona C; Agabiti N; Fusco D; Michelozzi P; Perucci CA. (2003). Air pollution and myocardial
infarction in Rome: a case-crossover analysis. Epidemiology 14: 528-535.
Dockery DW; Brunekreef B. (1996). Longitudinal studies of air pollution effects on lung function. American journal of
respiratory and critical care medicine 154: S250-256.
Dockery DW; Luttmann-Gibson H; Rich DQ; Link MS; Mittleman MA; Gold DR; Koutrakis P; Schwartz JD; Verrier RL.
(2005a). Association of air pollution with increased incidence of ventricular tachyarrhythmias recorded by implanted
cardioverter defibrillators. Environ Health Perspect 113: 670-674.
Dockery DW; Luttmann-Gibson H; Rich DQ; Link MS; Schwartz JD; Gold DR; Koutrakis P; Verrier RL; Mittleman MA.
(2005b). Particulate air pollution and nonfatal cardiac events. Part II. Association of air pollution with confirmed
arrhythmias recorded by implanted defibrillators. Res Rep Health Eff Inst: 83-126; discussion 127-148.
Dockery DW; Pope CA, 3rd; XuX; Spengler JD; Ware JH; Fay ME; Ferris BG, Jr.; SpeizerFE. (1993). An association between
air pollution and mortality in six U.S. cities. The New England journal of medicine 329: 1753-1759.
Dominici F; Dermott AMC; T.J. H. (2004a). Improved Semiparametric Time Series Models of Air Pollution and Mortality.
Journal of the American Statistical Association 99: 938-948.
Dominici F; McDermott A; Daniels M; Zeger SL; Samet J. (2003a). Revised Analyses of Time-Series Studies of Air Pollution
and Health: Mortality Among Residents of 90 Cities. Health Effects Institute, Boston, MA.
Dominici F; McDermott A; Zeger SL; Samet JM. (2002). On the use of generalized additive models in time-series studies of air
pollution and health. Am J Epidemiol 156: 193-203.
Dominici F; McDermott A; Zeger SL; Samet JM. (2003b). National maps of the effects of particulate matter on mortality:
exploring geographical variation. Environ Health Perspect 111: 39-44.
Dominici F; Peng RD; Bell ML; Pham L; McDermott A; Zeger SL; Samet JMCEH. (2006). Fine particulate air pollution and
hospital admission for cardiovascular and respiratory diseases. JAMA 295: 1127-1134.
Dominici F; Peng RD; Ebisu K; Zeger SL; Samet JM; Bell ML. (2007a). Does the effect of PM10 on mortality depend on PM
nickel and vanadium content? A reanalysis of theNMMAPS data. Environ Health Perspect 115: 1701-1703.
Dominici F; Peng RD; Zeger SL; White RH; Samet JM. (2007b). Particulate air pollution and mortality in the United States: did
the risks change from 1987 to 2000? Am J Epidemiol 166: 880-888.
Dominici F; Zanobetti A; Zeger SL; Schwartz J; Samet JM. (2004b). Hierarchical bivariate time series models: a combined
analysis of the effects of particulate matter on morbidity and mortality. Biostatistics 5: 341-360.
Dominici F; Zeger SL; Samet JM. (2000). A measurement error model for time-series studies of air pollution and mortality.
Biostatistics 1: 157-175.
Donahue WF; Allen EW; Schindler DW. (2006). Impacts of coal-fired power plants on trace metals and poly cyclic aromatic
hydrocarbons (PAHs) in lake sediments in central Alberta, Canada. Journal of Paleolimnology 35: 111-128.
Donaldson K; Stone V; Borm PJ; Jimenez LA; Gilmour PS; Schins RP; Knaapen AM; Rahman I; Faux SP; Brown DM; MacNee
W. (2003). Oxidative stress and calcium signaling in the adverse effects of environmental particles (PM10). Free Radic
Biol Med 34: 1369-1382.
Dorman DC; McManus BE; Parkinson CU; Manuel CA; McElveen AM; Everitt JI. (2004). Nasal Toxicity of Manganese Sulfate
and Manganese Phosphate in Young Male Rats Following Subchronic (13-Week) Inhalation Exposure. Inhalation
Toxicology 16: 481-488.
Dorman DC; Struve MF; James RA; Marshall MW; Parkinson CU; Wong BA. (2001). Influence of Particle Solubility on the
Delivery of Inhaled Manganese to the Rat Brain: Manganese Sulfate and Manganese Tetroxide Pharmacokinetics
Following Repeated (14-Day) Exposure. Toxicology and applied pharmacology 170: 79-87.
Dostert C; Petrilli V; Van Bruggen R; Steele C; Mossman BT; Tschopp J. (2008). Innate Immune Activation Through Nalp3
Inflammasome Sensing of Asbestos and Silica. Science 320: 674.
Downs SH; Schindler C; Liu LJ; Keidel D; Bayer-Oglesby L; Brutsche MH; Gerbase MW; Keller R; Kunzli N; Leuenberger P;
Probst-Hensch NM; Tschopp JM; Zellweger JP; Rochat T; Schwartz J; Ackermann-Liebrich U. (2007). Reduced
exposure to PM10 and attenuated age-related decline in lung function. The New England journal of medicine 357: 2338-
2347.
Drewnick F; Husain L; Diamond D; Weber R; Demeijian KL; Schwab JJ; Hogrefe O; Peters S. (2003). Intercomparison and
evaluation of four semi-continuous PM2.5 sulfate instruments. Atmospheric Environment 37: 3335-3350.
Drewnick F; Jayne JT; Canagaratna M; Worsnop DR; Demeijian KL. (2004b). Measurement of Ambient Aerosol Composition
During the PMTACS-NY 2001 Using an Aerosol Mass Spectrometer. Part II: Chemically Speciated Mass Distributions.
Aerosol Science and Technology 38: 104-117.
Drewnick F; Schwab J; Jayne J; Canagaratna M; Worsnop D; Demeijian K. (2004a). Measurement of Ambient Aerosol
Composition During the PMTACS-NY 2001 Using an Aerosol Mass Spectrometer. Part I: Mass Concentrations. Aerosol
Science and Technology 38: 92-103.
Dubowsky SD; Suh H; Schwartz J; Coull BA; Gold DR. (2006). Diabetes, obesity, and hypertension may enhance associations
between air pollution and markers of systemic inflammation. Environ Health Perspect 114: 992-998.
December 2008
R-18
DRAFT-DO NOT QUOTE OR CITE
-------
Dugandzic R; Dodds L; Stieb D; Smith-Doiron M. (2006). The association between low level exposures to ambient air pollution
and term low birth weight: a retrospective cohort study. Environ Health 5:3.
Duncan Fairlie T; Jacob DJ; Park RJ. (2007). The impact of transpacific transport of mineral dust in the United States.
Atmospheric Environment 41: 1251-1266.
Duriscoe DM; Luginbuhl CB; Moore CA. (2007). Measuring Night-Sky Brightness with a Wide-Field CCD Camera.
Publications of the Astronomical Society of the Pacific 119: 192-213.
Dusek U; Frank G; Flildebrandt L; Curtius J; Schneider J; Walter S; Chand D; Drewnick F; Flings S; Jung D; Borrmann S;
AndreaeM. (2006). Size matters more than chemistry for cloud-nucleating ability of aerosol particles. Science312: 1375-
1378.
Duvall RM; Norris GA; Dailey LA; Burke JM; McGee JK; Gilmour MI; Gordon T; Devlin RB. (2008). Source Apportionment of
Particulate Matter in the U.S. and Associations with Lung Inflammatory Markers. Inhalation Toxicology 20: 671 - 683.
Dvonch JT; Brook RD; Keeler GJ; Rajagopalan S; DAlecy LG; Marsik FJ; Morishita M; Yip FY; Brook JR; Timm EJ; Wagner
JG; Harkema JR. (2004). Effects of concentrated fine ambient particles on rat plasma levels of asymmetric
dimethylarginine. Inhalation Toxicology 16: 473-480.
Easter RC; Ghan SJ; Zhang Y; Say lor RD; Chapman EG; Laulainen NS; Abdul-Razzak H; Leung LR; Bian X; Zaveri RA.
(2004). MIRAGE: Model description and evaluation of aerosols and trace gases. Journal of Geophysical Research 109,
D20210, doi: 10.1029/2004JD004571.
Ebelt ST; Wilson WE; Brauer M. (2005). Exposure to ambient and nonambient components of particulate matter: a comparison
of health effects. Epidemiology 16: 396-405.
Eder B; Yu S. (2006). A performance evaluation of the 2004 release of Models-3 CMAQ. Atmospheric Environment 40: 4811-
4824.
Edetsberger M; Gaubitzer E; Valic E; Waigmann E; Kohler G. (2005). Detection of nanometer-sized particles in living cells
using modern fluorescence fluctuation methods. Biochemical and Biophysical Research Communications 332: 109-116.
Edgerton ES; Hartsell BE; Jansen JJ; Hansen DA; Waid CJ; Kandasamy K. (2004). 5-Year Trend Analysis of PM2.5 Data from
the SEARCH Network. Paper presented at the Regional and Global Perspectives on Haze: Causes, Consequences and
Controversies, Visibility Specialty Conference, Air and Waste Management Association.
Edney EO; Lewandowski M; Offenberg JH; Wang W; Claeys M; Kleindienst TE; Jaoui M. (2005). Formation of 2-methyl tetrols
and 2-methylglyceric acid in secondary organic aerosol from laboratory irradiated isoprene/NOX/SO 2/air mixtures and
their detection in ambient PM2.5 samples collected in the eastern United States. Atmospheric Environment 39: 5281-
5289.
Eeva T; Lehikoinen E. (2004). Rich calcium availability diminishes heavy metal toxicity in Pied Flycatcher. Functional Ecology
18: 548-553.
Eeva T; Tanhuanpaa S; Rabergh C; Airaksinen S; Nikinmaa M; Lehikoinen E. (2000). Biomarkers and fluctuating asymmetry as
indicators of pollution- induced stress in two hole-nesting passerines. Functional Ecology 14: 235-243.
Eftim SE; Samet JM; Janes H; McDermott A; Dominici F. (2008). Fine particulate matter and mortality: a comparison of the six
cities and American Cancer Society cohorts with a medicare cohort. Epidemiology 19: 209-216.
Eiguren-Fernandez A; Miguel AH; Jaques PA; Sioutas C. (2003). Evaluation of a Denuder-MOUDI-PUF Sampling System to
Measure the Size Distribution of Semi-Volatile Polycyclic Aromatic Hydrocarbons in the Atmosphere. Aerosol Science
and Technology 37: 201-209.
Elder A; Gelein R; Finkelstein J; Phipps R; Frampton M; Utell M; Kittelson DB; Watts WF; Hopke P; Jeong CH; Kim E; Liu W;
Zhao W; Zhuo L; Vincent R; Kumarathasan P; Oberdorster G. (2004a). On-road exposure to highway aerosols. 2.
Exposures of aged, compromised rats. Inhalation Toxicology 16 Suppl 1: 41-53.
Elder A; Gelein R; Silva V; Feikert T; Opanashuk L; Carter J; Potter R; Maynard A; Ito Y; Finkelstein J; Oberdorster G. (2006).
Translocation of inhaled ultrafme manganese oxide particles to the central nervous system. Environ Health Perspect 114:
1172-1178.
Elder AC; Gelein R; Azadniv M; Frampton M; Finkelstein J; Oberdorster G. (2004b). Systemic effects of inhaled ultrafme
particles in two compromised, aged rat strains. Inhalation Toxicology 16: 461-471.
Elliott SJ; Cole DC; Krueger P; Voorberg N; Wakefield S. (1999). The Power of Perception: Health Risk Attributed to Air
Pollution in an Urban Industrial Neighbourhood. Risk Analysis 19: 621-634.
Ely DW; Leary JT; Stewart TR; Ross DM. (1993). The Establishment of the Denver Visibility Standard.
El-Zanan HS; Lowenthal DH; Zielinska B; Chow JC; Kumar N. (2005). Determination of the organic aerosol mass to organic
carbon ratio in IMPROVE samples. Chemosphere 60: 485-496.
Engel-Cox JA; Weber SA. (2007). Compilation and assessment of recent positive matrix factorization and UNMIX receptor
model studies on fine particulate matter source apportionment for the eastern United States. J Air Waste Manag Assoc 57:
1307-1316.
Engewald W; Teske J; Efer J. (1999). Programmed temperature vaporisers-based large volume injection in capillary gas
chromatography. Journal of Chromatography A 842: 143-161.
England GC; Watson JG; Chow JC; Zielinska B; Chang M. (2007b). Dilution-Based Emissions Sampling from Stationary
Sources: Part 2- Gas-Fired Combustors Compared with Other Fuel-Fired Systems. Journal of the Air & Waste
Management Association 57.
December 2008
R-19
DRAFT-DO NOT QUOTE OR CITE
-------
England GC; Watson JG; Chow JC; Zielinska B; Oliver Chang MC; Loos KR; Hidv GM. (2007a). Dilution-based emissions
sampling from stationary sources: Part 1-Compact sampler methodology and performance. Journal of the Air & Waste
Management Association^ 995) 57: 65-78.
Englert N. (2004). Fine particles and human health - a review of epidemiological studies. Toxicology Letters 149: 235-242.
Enna SJ; Schanker LS. (1972). Absorption of drugs from the rat lung. American Journal of Physiology 223: 1227.
Enstrom JE. (2005). Fine particulate air pollution and total mortality among elderly Californians, 1973-2002. Inhal Toxicol 17:
803-816.
Erbas B; Kelly AM; Physick B; Code C; Edwards M. (2005). Air pollution and childhood asthma emergency hospital
admissions: estimating intra-city regional variations. Int J Environ Health Res 15: 11-20.
Erbel R; Mohlenkamp S; Kerkhoff G; Budde T; Schmermund A. (2007). Non-invasive screening for coronary artery disease:
calcium scoring. Heart 93: 1620-1629.
Erdinger T; Durr M; Hopker KA. (2005). Correlations between mutagenic activity of organic extracts of airborne particulate
matter, NOx and sulphur dioxide in southern Germany: results of a two-year study. Environ Sci Pollut Res Int 12: 10-20.
Ericksen JA; Gustin MS; Schorran DE; Johnson DW; Lindberg SE; Coleman JS. (2003). Accumulation of atmospheric mercury
in forest foliage. Atmospheric Environment 37: 1613-1622.
Erisman J-W; Vermetten AWM; Asman WAH; Waijers-IJpelaan A; Slanina J. (1988). Vertical distribution of gases and aerosols:
the behaviour of ammonia and related components in the lower atmosphere. Atmospheric Environment 22: 1153-1160.
Escribano R; Sloan JJ; Siddique N; Sze N; Dudev T. (2001). Raman spectroscopy of carbon-containing particles. Vibrational
Spectroscopy 26: 179-186.
Evans GW; Jacobs SV. (1982). Air Pollution and Human Behavior. In Evans G (Ed.), Environmental Stress (pp. 105-132). New
York, NY: Cambridge University Press.
FalkR; PhilipsonK; Svartengren M; JarvisN; Bailey M; CamnerP. (1997). Clearance ofParticles from Small Ciliated Airways.
Experimental lung research 23: 495-515.
Fan X; Brook JR; Mabury SA. (2003). Sampling Atmospheric Carbonaceous Aerosols Using an Integrated Organic Gas and
Particle Sampler. Environmental science & technology 37: 3145-3151.
Fan X; Lee PKH; Brook JR; Mabury SA. (2004). Improved Measurement of Seasonal and Diurnal Differences in the
Carbonaceous Components of Urban Particulate Matter Using a Denuder-Based Air Sampler. Aerosol Science and
Technology 38: 63-69.
Fan Y; Zhang FS; Zhu J; Liu Z. (2008). Effective utilization of waste ash from MSW and coal co-combustion power plant:
Zeolite synthesis. J Hazard Mater 153: 382-388.
Farkas A; Balashazy I. (2008). Quantification of particle deposition in asymmetrical tracheobronchial model geometry. Comput
Biol Med 38: 508-518.
Farkas A; Balashazy I; SzQcs K. (2006). Characterization of Regional and Local Deposition of Inhaled Aerosol Drugs in the
Respiratory System by Computational Fluid and Particle Dynamics Methods. J Aerosol Med 19: 329-343.
Farmer PB; Singh R; Kaur B; Sram RJ; Binkova B; Kalina I; Popov TA; Garte S; Taioli E; Gabelova A; Cebulska-Wasilewska
A. (2003). Molecular epidemiology studies of carcinogenic environmental pollutants. Effects of polycyclic aromatic
hydrocarbons (PAHs) in environmental pollution on exogenous and oxidative DNA damage. MutatRes 544: 397-402.
Farraj AK; Haykal-Coates N; Ledbetter AD; Evansky PA; Gavett SH. (2006a). Inhibition of pan neurotrophin receptor p75
attenuates diesel particulate-induced enhancement of allergic airway responses in C57/B16J mice. Inhalation toxicology
18: 483-491.
Farraj AK; Haykal-Coates N; Ledbetter AD; Evansky PA; Gavett SH. (2006b). Neurotrophin mediation of allergic airways
responses to inhaled diesel particles in mice. Toxicological sciences 94: 183-192.
Fedulov AV; Leme A; Yang Z; Dahl M; Lim R; Mariani TJ; Kobzik L. (2008). Pulmonary exposure to particles during
pregnancy causes increased neonatal asthma susceptibility. American journal of respiratory cell and molecular biology
38: 57-67.
Feichter J; Sausen R; GraBl H; Fiebig M. (2003). Comment on "Control of fossil-fuel particulate black carbon and organic
matter, possibly the most effective method of slowing global warming" by M. Z. Jacobson. Journal of Geophysical
Research 108 (D24), 4767, doi:10.1029/2002JD003223.
Feldpausch P; Fiebig M; Fritzsche L; Petzold A. (2006). Measurement of ultrafine aerosol size distributions by a combination of
diffusion screen separators and condensation particle counters. J Aerosol Sci 37: 577-597.
Feng MH; Shan XQ; Zhang SZ; Wen B. (2005). Comparison of a rhizosphere-based method with other one-step extraction
methods for assessing the bioavailability of soil metals to wheat. Chemosphere 59: 939-949.
Ferdinands JM; Crawford CA; Greenwald R; Van Sickle D; Hunter E; Teague WG. (2008). Breath acidification in adolescent
runners exposed to atmospheric pollution: a prospective, repeated measures observational study. Environ Health 7: 10.
Ferin J; Oberdorster G; Penney DP. (1992). Pulmonary retention of ultrafine and fine particles in rats. Am J Respir Cell Mol Biol
6: 535-542.
Ferm M; Watt J; O'Hanlon S; De Santis F; Varotsos C. (2006). Deposition measurement of particulate matter in connection with
corrosion studies. Analytical and Bioanalytical Chemistry 384: 1320-1330.
Filleul L; Baldi I; Dartigues JF; Tessier JF. (2003). Risk factors among elderly for short term deaths related to high levels of air
pollution. Occupational and environmental medicine 60: 684-688.
Filleul L; Le Tertre A; Baldi I; Tessier JF. (2004). Difference in the relation between daily mortality and air pollution among
elderly and all-ages populations in southwestern France. Environ Res 94: 249-253.
December 2008
R-20
DRAFT-DO NOT QUOTE OR CITE
-------
Filleul L; Rondeau V; Vandentorren S; Le Moual N; Cantagrel A; Annesi-Maesano I; Charpin D; Declercq C; Neukirch F; Paris
C; Vervloet D; Brochard P; Tessier JF; Kauffmann F; Baldi I. (2005). Twenty five year mortality and air pollution:
results from the French PAARC survey. Occupational and environmental medicine 62: 453-460.
Finch GL; Hobbs CH; Blair LF; Barr EB; Hahn FF; Jaramillo RJ; Kubatko JE; March TH; White RK; Krone JR; Menache MG;
Nikula KJ; Mauderly JL; Van Gerpen J; Merceica MD; Zielinska B; Stankowski L; Burling K; Howell S. (2002). Effects
of subchronic inhalation exposure of rats to emissions from a diesel engine burning soybean oil-derived biodiesel fuel.
Inhal Toxicol 14: 1017-1048.
Fine PM; Chakrabarti B; Krudysz M; Schauer JJ; Sioutas C. (2004). Diurnal Variations of Individual Organic Compound
Constituents of Ultrafme and Accumulation Mode Particulate Matter in the Los Angeles Basin. Environmental science &
technology 38: 1296-1304.
Fine PM; Jaques PA; Flering SV; Sioutas C. (2003). Performance Evaluation and Use of a Continuous Monitor for Measuring
Size-Fractionated PM2.5 Nitrate. Aerosol Science and Technology 37: 342 - 354.
Finkelstein MM; Jerrett M; DeLuca P; Finkelstein N; Verma DK; Chapman K; Sears MR. (2003). Relation between income, air
pollution and mortality: a cohort study. CMAJ 169: 397-402.
Finlay WH; Martin AR. Recent Advances in Predictive Understanding of Respiratory Tract Deposition. J Aerosol Med: 1-18.
Finlay son-Pitts BJ; Pitts JN. (2000). Chemistry of the Upper and Lower Atmosphere: Theory, Experiments, and Applications:
Academic Press.
Finnerty K; Choi J-E; Lau A; Davis-Gorman G; Diven C; Seaver N; Linak William P; Witten M; McDonagh Paul F. (2007).
Instillation of coarse ash particulate matter and lipopolysaccharide produces a systemic inflammatory response in mice.
Journal of toxicology and environmental health Part A 70: 1957-1966.
Fiore AM; Jacob DJ; Mathur R; Martin RV. (2003). Application of empirical orthogonal functions to evaluate ozone simulations
with regional and global models. J Geophys Res 108: 10.1029.
Fischer PH; Steerenberg PA; Snelder JD; van Loveren H; van Amsterdam JG. (2002). Association between exhaled nitric oxide,
ambient air pollution and respiratory health in school children. Int Arch Occup Environ Health 75: 348-353.
Fishman AP; Elias JA. (1980). Fishman's Pulmonary diseases and disorders. New York, NY: McGraw-Hill.
Fismes J; Perrin-Ganier C; Empereur-Bissonnet P; Morel JL. (2002). Soil-to-root transfer and translocation of poly cyclic
aromatic hydrocarbons by vegetables grown on industrial contaminated soils. J Environ Qual 31: 1649-1656.
Forastiere F; Faustini A. (2008). Are we understanding the respiratory effects of traffic related airborne particles? Thorax 63:
574-576.
Forastiere F; Stafoggia M; Picciotto S; Bellander T; DTppoliti D; Lanki T; von Klot S; Nyberg F; Paatero P; Peters A; Pekkanen
J; Sunyer J; Perucci CA. (2005). A case-crossover analysis of out-of-hospital coronary deaths and air pollution in Rome,
Italy. American journal of respiratory and critical care medicine 172: 1549-1555.
Forman JP; Brenner BM. (2006). 'Hypertension' and 'microalbuminuria': the bell tolls for thee. Kidney Int 69: 22-28.
Forsberg B, Sternberg N, Linne R, Segerstedt B, Wall S. 1998. Daily air pollution levels and acute asthma in
southern Sweden. Eur Respir J 12(4): 900-905.
Forster P; Ramaswamy V; Artaxo P; Berntsen T; Betts R; Fahey DW; Haywood J; Lean J; Lowe DC; Myhre G. (2007). Changes
in Atmospheric Constituents and in Radiative Forcing. Climate Change: 129-234.
Foster PMD; Mylchreest E; Gaido KW; Sar M. (2001). Effects of phthalate esters on the developing reproductive tract of male
rats. Human Reproduction Update 7: 231-235.
Foster PMD; Thomas LV; Cook MW; Gangolli SD. (1980). Study of the testicular effects and changes in zinc excretion
produced by some n-alkyl phthalates in the rat. Toxicology and applied pharmacology 54: 392-398.
Foster WM; Stetkiewicz PT. (1996). Regional clearance of solute from the respiratory epithelia: 18-20 h postexposure to ozone.
Journal of applied physiology 81: 1143-1149.
Foster WM; Wagner EM. (2001). Bronchial edema alters 99mTc-DTPA clearance from the airway surface in sheep. Journal of
applied physiology 91: 2567-2573.
Fowler D; Cape JN; Unsworth MH. (1989). Deposition of atmospheric pollutants on forests. Philosophical Transactions of the
Royal Society of London B 324: 247-265.
Fowler D; Duyzer JH; Baldocchi DD. (1991). Inputs of trace gases, particles and cloud droplets to terrestrial surfaces.
Proceedings of the Royal Society of Edinburgh Sect B: Biol Sci 97: 35-59.
Fox GA. (1991). Practical causal inference for ecoepidemiologists. J Toxicol Environ Health 33: 359-373.
Frampton MW. (2001). Systemic and cardiovascular effects of airway injury and inflammation: Ultrafme particle exposure in
humans. Environmental Health Perspectives 109: 529-532.
Frampton MW; Stewart JC; Oberdorster G; Morrow PE; Chalupa D; Pietropaoli AP; Frasier LM; Speers DM; Cox C; Huang LS;
Utell MJ. (2006). Inhalation of ultrafine particles alters blood leukocyte expression of adhesion molecules in humans.
Environmental Health Perspectives 114: 51-58.
Frank NH. (2006). Retained nitrate, hydrated sulfates, and carbonaceous mass in federal reference method fine particulate matter
for six eastern US cities. J Air Waste Manag Assoc 56: 500-511.
Franklin M; Koutrakis P; Schwartz P. (2008). The role of particle composition on the association between PM2.5 and mortality.
Epidemiology 19: 680-689.
Franklin M; Schwartz J. (2008). The Impact of Secondary Particles on the Association between Ambient Ozone and Mortality.
Environ Health Perspect 116: 453-458.
December 2008
R-21
DRAFT-DO NOT QUOTE OR CITE
-------
Franklin M; Zeka A; Schwartz J. (2007). Association between PM2.5 and all-cause and specific-cause mortality in 27 US
communities. Journal of Exposure Science and Environmental Epidemiology 17: 279-287.
Franklin SS; Sutton-Tyrrell K; Belle SH; Weber MA; Kuller LEI. (1997). The importance of pulsatile components of
hypertension in predicting carotid stenosis in older adults. J Elypertens 15:1143-1150.
Fraser MP; Cass GR; Simoneit BRT. (1999). Particulate organic compounds emitted from motor vehicle exhaust and in the urban
atmosphere. Atmospheric Environment 33: 2715-2724.
Fraser MP; Yue ZW; Buzcu B. (2003). Source apportionment of fine particulate matter in Houston, TX, using organic molecular
markers. Atmospheric Environment 37: 2117-2123.
Freer-Smith PE1; El-khatib A; Taylor G. (2004). Capture of particulate pollution by trees: a comparison of species typical of
semi-arid areas (Ficus nitida and Eucalyptus globulus) with European and North American species. Water Air and Soil
Pollution 155: 173-187.
Fritze H; Niini S; Mikkola K; Makinen A. (1989). Soil microbial effects of a Cu-Ni smelter in southwestern Finland. Biology and
Fertility of Soils 8: 87-94.
Fruin S; Westerdahl D; Sax T; Sioutas C; Fine PM. (2008). Measurements and predictors of on-road ultrafine particle
concentrations and associated pollutants in Los Angeles. Atmospheric Environment 42: 207-219.
Fryer ME; Collins CD. (2003). Model intercomparison for the uptake of organic chemicals by plants. Environ Sci Technol 37:
1617-1624.
Fuentes M; Raftery AE. (2005). Model Evaluation and Spatial Interpolation by Bayesian Combination of Observations with
Outputs from Numerical Models. Biometrics 61: 36-45.
Fujimaki H; Kurokawa Y. (2004). Diesel exhaust-associated gas components enhance chemokine production by cervical lymph-
node cells from mice immunized with sugi basic proteins. Inhalation toxicology 16: 61-65.
Fujimaki H; Kurokawa Y; Yamamoto S; Satoh M. (2006). Distinct requirements for interleukin-6 in airway inflammation
induced by diesel exhaust in mice. Immunopharmacology and immunotoxicology 28: 703-714.
Fujimaki H; Yamamoto S; Kurokawa Y. (2005). Effect of diesel exhaust on immune responses in C57BL/6 mice intranasally
immunized with pollen antigen. Journal of UOEH 27: 11-24.
Fujimoto A; Tsukue N; Watanabe M; Sugawara I; Yanagisawa R; Takano H; Yoshida S; Takeda K. (2005). Diesel exhaust
affects immunological action in the placentas of mice. Environmental toxicology 20: 431-440.
Fung KY; Khan S; Krewski D; Chen Y. (2006). Association between air pollution and multiple respiratory hospitalizations
among the elderly in Vancouver, Canada. Inhal Toxicol 18: 1005-1011.
Fung KY; Luginaah I; Gorey KM; Webster G. (2005b). Air pollution and daily hospital admissions for cardiovascular diseases in
Windsor, Ontario. Can J Public Health 96: 29-33.
Fung KY; Luginaah I; K.M. G; Webster G. (2005a). Air Pollution and Daily Hospitalization Rates for Cardiovascular and
Respiratory Diseases in London, Ontario. International Journal of Environmental Studies 62: 677-685.
Gabelova A; Valovicova Z; Bacova G; Labaj J; Binkova B; Topinka J; Sevastyanova O; Sram RJ; Kalina I; Habalova V; Popov
TA; Panev T; Farmer PB. (2007a). Sensitivity of different endpoints for in vitro measurement of genotoxicity of
extractable organic matter associated with ambient airborne particles (PM10). Mutat Res 620: 103-113.
Gabelova A; Valovicova Z; Labaj J; Bacova G; Binkova B; Farmer Peter B. (2007b). Assessment of oxidative DNA damage
formation by organic complex mixtures from airborne particles PM(10). Mutation research 620: 135-144.
Galan I; Tobias A; Banegas JR; Aranguez E. (2003). Short-term effects of air pollution on daily asthma emergency room
admissions. Eur Respir J 22: 802-808.
Gallagher J; Sams R; Inmon J; Gelein R; Elder A; Oberdorster G; Prahalad AK. (2003). Formation of 8-oxo-7,8-dihydro-2'-
deoxyguanosine in rat lung DNA following subchronic inhalation of carbon black. Toxicology and applied pharmacology
190: 224-231.
Gallagher MW; Choularton TW; Morse AP; Fowler D. (1988). Measurements of the size dependence of cloud droplet deposition
at a hill site. Quarterly Journal of the Royal Meteorological Society 114: 1291-1303.
Gandhi N; Bhavsar SP; Diamond ML; Kuwabara JS; Marvin-DiPasquale M; Krabbenhoft DP. (2007). Development of a mercury
speciation, fate, and biotic uptake (BIOTRANSPEC) model: Application to Lahontan Reservoir (Nevada, USA).
Environmental Toxicology and Chemistry 26: 2260-2273.
Gao F; Barchowsky A; Nemec Antonia A; Fabisiak James P. (2004a). Microbial stimulation by Mycoplasma fermentans
synergistically amplifies IL-6 release by human lung fibroblasts in response to residual oil fly ash (ROFA) and nickel.
Toxicological sciences 81: 467-479.
Gao S; Ng NL; Keywood M; Varutbangkul V; Bahreini R; Nenes A; He J; Yoo KY; Beauchamp JL; Hodyss RP; Flagan RC;
Seinfeld JH. (2004b). Particle phase acidity and oligomer formation in secondary organic aerosol. Environ Sci Technol
38: 6582-6589.
Gao YZ; Zhu LZ. (2004). Plant uptake, accumulation and translocation of phenanthrene and pyrene in soils. Chemosphere 55:
1169-1178.
Garner JHB; Pagano T; Cowling EB. (1989). An evaluation of the role of ozone, acid deposition, and other airborne pollutants in
the forests of eastern North America. U.S. Department of Agriculture, Forest Service, Southeastern Forest Experiment
Station, Asheville, NC. General Technical Report SE-59.
Garten CT, Jr.; Hanson PJ. (1990). Foliar retention of 15N-nitrate and 15N-ammonium by red maple (Acer rubrum) and white
oak (Quercus alba) leaves from simulated rain. Environmental and Experimental Botany 30: 333-342.
December 2008
R-22
DRAFT-DO NOT QUOTE OR CITE
-------
Gauderman WJ; Avol E; Gilliland F; Vora H; Thomas D; Berhane K; McConnell R; Kuenzli N; Lurmann F; Rappaport E;
Margolis H; Bates D; Peters J. (2004). The effect of air pollution on lung development from 10 to 18 years of age. The
New England journal of medicine 351: 1057-1067.
Gauderman WJ; Gilliland GF; Vora H; Avol E; Stram D; McConnell R; Thomas D; Lurmann F; Margolis FIG; Rappaport EB;
Berhane K; Peters JM. (2002). Association between air pollution and lung function growth in southern California
children: results from a second cohort. American journal of respiratory and critical care medicine 166: 76-84.
Gauderman WJ; McConnell R; Gilliland F; London S; Thomas D; Avol E; Vora H; Berhane K; Rappaport EB; Lurmann F;
Margolis FIG; Peters J. (2000). Association between air pollution and lung function growth in southern California
children. American journal of respiratory and critical care medicine 162: 1383-1390.
Gavett SH; Flaykal-Coates N; Copeland LB; Fleinrich J; Gilmour MI. (2003). Metal composition of ambient PM2.5 influences
severity of allergic airways disease in mice. Environmental health perspectives 111: 1471-1477.
Gawel JE; Ahner BA; Friedland AJ; Morel FMM. (1996). Role for heavy metals in forest decline indicated by phytochelatin
measurements. Nature 381: 4-249.
Gbor PK; Wen D; Meng F; Yang F; Sloan JJ. (2007). Modeling of mercury emission, transport and deposition in North America.
Atmospheric Environment 41: 1135-1149.
Gee GC; Payne-Sturges DC. (2004). Environmental health disparities: a framework integrating psychosocial and environmental
concepts. Environ Flealth Perspect 112: 1645-1653.
Gehlbach BK; Geppert E. (2004). The pulmonary manifestations of left heart failure. Chest 125: 669-682.
Gehrig R; Buchmann B. (2003). Characterising seasonal variations and spatial distribution of ambient PM10 and PM2. 5
concentrations based on long-term Swiss monitoring data. Atmospheric Environment 37: 2571-2580.
Gehring U; Cyrys J; Sedlmeir G; Brunekreef B; Bellander T; Fischer P; Bauer CP; Reinhardt D; Wichmann FIE; Fleinrich J.
(2002). Traffic-related air pollution and respiratory health during the first 2 yrs of life. Eur Respir J 19: 690-698.
Gehring U; Fleinrich J; Kramer U; Grote V; Flochadel M; Sugiri D; Kraft M; Rauchfuss K; Eberwein FIG; Wichmann FIE.
(2006). Long-term exposure to ambient air pollution and cardiopulmonary mortality in women. Epidemiology 17: 545-
551.
Geiser M; Rothen-Rutishauser B; Kapp N; Schurch S; Kreyling W; Schulz H; Semmler M; Im Hof V; Fleyder J; Gehr P. (2005).
Ultrafme particles cross cellular membranes by nonphagocytic mechanisms in lungs and in cultured cells. Environmental
health perspectives 113: 1555-1560.
Gelencser A; Meszaros T; Blazso M; Kiss G; Krivacsy Z; Molnar A; Meszaros E. (2000). Structural Characterisation of Organic
Matter in Fine Tropospheric Aerosol by Pyrolysis-Gas Chromatography-Mass Spectrometry. J Atmos Chem 37: 173-183.
Gelencser A; Varga Z. (2005). Evaluation of the atmospheric significance of multiphase reactions in atmospheric secondary
organic aerosol formation. Atmos Chem Phys 5: 2823-2831.
Geller MD; Ntziachristos L; Mamakos A; Samaras Z; Schmitz DA; Froines JR; Sioutas C. (2006). Physicochemical and redox
characteristics of particulate matter (PM) emitted from gasoline and diesel passenger cars. Atmospheric Environment 40:
6988-7004.
Generoso S; Bey I; Attie J-L; Breon F-M. (2007). A satellite- and model-based assessment of the 2003 Russian fires: impact on
the arctic region. Journal of Geophysical Research 112: 5302.
Geng H; Meng Z; Zhang Q. (2005). Effects of blowing sand fine particles on plasma membrane permeability and fluidity, and
intracellular calcium levels of rat alveolar macrophages. Toxicology letters 157: 129-137.
Geng H; Meng Z; Zhang Q. (2006). In vitro responses of rat alveolar macrophages to particle suspensions and water-soluble
components of dust storm PM(2.5). Toxicology in vitro 20: 575-584.
Gent JF; Triche EW; Holford TR; Belanger K; Bracken MB; Beckett WS; Leaderer BP. (2003). Association of low-level ozone
and fine particles with respiratory symptoms in children with asthma. JAMA 290: 1859-1867.
Georgopoulos PG; Wang SW; Vyas VM; Sun Q; Burke J; Vedantham R; McCurdy T; Ozkaynak H. (2005). A source-to-dose
assessment of population exposures to fine PM and ozone in Philadelphia, PA, during a summer 1999 episode. Journal of
exposure analysis and environmental epidemiology 15: 439-457.
Gerlofs-Nijland ME; Dormans JA; Bloemen HJ; Leseman DL; John A; Boere F; Kelly FJ; Mudway IS; Jimenez AA; Donaldson
K; Guastadisegni C; Janssen NA; Brunekreef B; Sandstrom T; van Bree L; Cassee FR. (2007). Toxicity of coarse and
fine particulate matter from sites with contrasting traffic profiles. Inhalation Toxicology 19: 1055-1069.
Geron C; Rasmussen R; Arnts RR; Guenther A. (2000). A review and synthesis of monoterpene speciation from forests in the
United States. Atmospheric Environment 34: 1761-1781.
Geroulakos G; O'Gorman DJ; Kalodiki E; Sheridan DJ; Nicolaides AN. (1994). The carotid intima-media thickness as a marker
of the presence of severe symptomatic coronary artery disease. European heart journal 15: 781-785.
Gerritsen J; Carlson RE; Dycus DL; Faulkner C; Gibson GR. (1998). Lake and reservoir bioassessment and biocriteria: Technical
guidance document. EPA 841-B-98-007. US Environmental Protection Agency, Office of Water, Washington, DC.
http://www.epa.gov/owow/monitoring/tech/lakes.html,
Gerstein HC; Mann JF; Yi Q; Zinman B; Dinneen SF; Hoogwerf B; Halle JP; Young J; Rashkow A; Joyce C; Nawaz S; Yusuf S.
(2001). Albuminuria and risk of cardiovascular events, death, and heart failure in diabetic and nondiabetic individuals.
JAMA 286: 421-426.
Geys J; Coenegrachts L; Vercammen J; Engelborghs Y; Nemmar A; Nemery B; Hoet PHM. (2006). In vitro study of the
pulmonary translocation of nanoparticles A preliminary study. Toxicology Letters 160: 218-226.
December 2008
R-23
DRAFT-DO NOT QUOTE OR CITE
-------
Ghelfi E; Rhoden CR; Wellenius GA; Lawrence J; Gonzalez-Flecha B. (2008). Cardiac oxidative stress and electrophysiological
changes in rats exposed to concentrated air particles are mediated by TRP-dependent pulmonary reflexes. Toxicol Sci
102: 328-336.
Ghio AJ; Churg A; Roggli VL. (2004). Review: Ferruginous Bodies: Implications in the Mechanism of Fiber and Particle
Toxicity. Toxicologic pathology 32: 643.
Ghio AJ; Cohen MD. (2005). Disruption of iron homeostasis as a mechanism of biologic effect by ambient air pollution particles.
Inhal Toxicol 17: 709-716.
Ghio AJ; Devlin RB. (2001). Inflammatory lung injury after bronchial instillation of air pollution particles. Am J Respir Crit Care
Med 164: 704-708.
Ghio A J; Hall A; Bassett MA; Cascio WE; Devlin RB. (2003). Exposure to concentrated ambient air particles alters hematologic
indices in humans. Inhalation Toxicology 15: 1465-1478.
Ghio AJ; Kim C; Devlin RB. (2000). Concentrated ambient air particles induce mild pulmonary inflammation in healthy human
volunteers. Am J Respir Crit Care Med 162: 981-988.
Ghosh R; Rankin J; Pless-Mulloli T; Glinianaia S. (2007). Does the effect of air pollution on pregnancy outcomes differ by
gender? A systematic review. Environ Res 105: 400-408.
Giglio L; van der Werf GR; Randerson JT; Collatz GJ; Kasibhatla P. (2006). Global estimation of burned area using MODIS
active fire observations. Atmos Chem Phys 6: 957-974.
Gilboa SM; Mendola P; Olsham AF; Langlois PH; Savitz DA; Loomis D; Herring AH; Fixler DE. (2005). Relation between
ambient air quality and selected birth defects, seven county study, Texas, 1997-2000. Am J Epidemiol 162: 238-252.
Gillette DA; Hanson KJ. (1989). Spatial and temporal variability of dust production caused by wind erosion in the United States.
J Geophys Res 94: 2197-2206.
Gilli G; Bono R; La RA; Traversi D; Pignata C; Schilir* T. (2007). The mutagenic hazards of environmental PM2.5 in Turin.
Environmental Research 103: 168-175.
Gilliam RC; Childs PP; Huber AH; Raman S. (2005). Metropolitan-scale Transport and Dispersion from the New York World
Trade Center Following September 11,2001. Parti: AnEvaluation of the CALMET Meteorological Model. Pure and
Applied Geophysics 162: 1981-2003.
Gilliland F; Avol PK; Jerrett M; Dvonch T; Lurmann F; Buckley T; Breysse P; Keeler G; de Villiers T; McConnell R. (2005).
Air Pollution Exposure Assessment for Epidemiologic Studies of Pregnant Women and Children: Lessons Learned from
the Centers for Children's Environmental Health and Disease Prevention Research. Environmental health perspectives
113: 1447.
Gilliland FD; Li YF; Saxon A; Diaz-Sanchez D. (2004). Effect of glutathione-S-transferase Ml and PI genotypes on xenobiotic
enhancement of allergic responses: randomised, placebo-controlled crossover study. The Lancet 363: 119-125.
Gilliland FD; McConnell R; Peters J; Gong Jr H. (1999). A Theoretical Basis for Investigating Ambient Air Pollution and
Children's Respiratory Health. Environmental health perspectives 107: 403-407.
Gilmour MI; McGee J; Duvall Rachelle M; Dailey L; Daniels M; Boykin E; Cho S-H; Doerfler D; Gordon T; Devlin Robert B.
(2007). Comparative toxicity of size-fractionated airborne particulate matter obtained from different cities in the United
States. Inhalation toxicology 19 Suppl 1: 7-16.
Gilmour MI; O'Connor S; Dick Colin AJ; Miller CA; Linak WP. (2004a). Differential pulmonary inflammation and in vitro
cytotoxicity of size-fractionated fly ash particles from pulverized coal combustion. Journal of the Air & Waste
Management Association 54: 286-295.
Gilmour PS; Morrison ER; Vickers MA; Ford I; Ludlam CA; Greaves M; Donaldson K; MacNee W. (2005). The procoagulant
potential of environmental particles (PM10). Occup Environ Med 62: 164-171.
Gilmour PS; Nyska A; Schladweiler MC; McGee JK; Wallenborn JG; Richards JH; Kodavanti UP. (2006a). Cardiovascular and
blood coagulative effects of pulmonary zinc exposure. Toxicol Appl Pharmacol 211: 41-52.
Gilmour PS; Rahman I; Donaldson K; MacNee W. (2003). Histone acetylation regulates epithelial IL-8 release mediated by
oxidative stress from environmental particles. Am J Physiol Lung Cell Mol Physiol 284: L533-540.
Gilmour PS; Schladweiler MC; Nyska A; McGee JK; Thomas R; Jaskot RH; Schmid J; Kodavanti UP. (2006b). Systemic
imbalance of essential metals and cardiac gene expression in rats following acute pulmonary zinc exposure. Journal of
toxicology and environmental health Part A 69: 2011-2032.
Gilmour PS; Schladweiler MC; Richards JH; Ledbetter AD; Kodavanti UP. (2004b). Hypertensive rats are susceptible to TLR4-
mediated signaling following exposure to combustion source particulate matter. Inhalation toxicology 16 Suppl 1: 5-18.
Gilmour PS; Ziesenis A; Morrison ER; Vickers MA; Drost EM; Ford I; Karg E; Mossa C; Schroeppel A; Ferron GA; Hey der J;
Greaves M; MacNee W; Donaldson K. (2004c). Pulmonary and systemic effects of short-term inhalation exposure to
ultrafme carbon black particles. Toxicol Appl Pharmacol 195: 35-44.
Girardot SP; Ryan PB; Smith SM; Davis WT; Hamilton CB; Obenour RA; Renfro JR; Tromatore KA; Reed GD. (2006). Ozone
and PM2.5 exposure and acute pulmonary health effects: a study of hikers in the Great Smoky Mountains National Park.
Environ Health Perspect 114: 1044-1052.
Girerd X; Mourad JJ; Acar C; Heudes D; Chiche S; Bruneval P; Mignot JP; Billaud E; Safar M; Laurent S. (1994). Noninvasive
measurement of medium-sized artery intima-media thickness in humans: in vitro validation. JVascRes31: 114-120.
Givati A; Rosenfeld D. (2004). Quantifying precipitation suppression due to air pollution. Journal of Applied Meteorology 43:
1038-1056.
December 2008
R-24
DRAFT-DO NOT QUOTE OR CITE
-------
Gleason SM; Faucette DT; Toyofuku MM; Torres CA; Bagley CF. (2007). Assessing and mitigating the effects of windblown
soil on rare and common vegetation. Environmental Management 40: 1016-1024.
Glinianaia SV; Rankin J; Bell R; Pless-Mulloli T; Howel D. (2004). Does particulate air pollution contribute to infant death?
Environ Health Perspect 112: 1365-1370.
Go AS; Elylek EM; Phillips KA; Chang Y; Elenault LE; Selby JV; Singer DE. (2001). Prevalence of Diagnosed Atrial Fibrillation
in Adults: National Implications for Rhythm Management and Stroke Prevention: the AnTicoagulation and Risk Factors
In Atrial Fibrillation (ATRIA) Study. Jama 285: 2370-2375.
Godleski JJ; Clarke RW; Coull BA; Saldiva PUN; Jiang NF; Lawrence J; Koutrakis P. (2002). Composition of inhaled urban air
particles determines acute pulmonary responses. Ann Occup Hyg 46: 419-424.
Godleski JJ; Verrier RL; Koutrakis P; Catalano P; Coull B; Reinisch U; Lovett EG; Lawrence J; Murthy GG; Wolfson JM;
Clarke RW; Nearing BD; Killingsworth C. (2000). Mechanisms of morbidity and mortality from exposure to ambient air
particles. Res Rep Health Eff Inst: 5-88; discussion 89-103.
Goforth MR; Christoforou CS. (2006). Particle size distribution and atmospheric metals measurements in a rural area in the
Southeastern USA. Science of the Total Environment 356: 217-227.
Gold DR; Litonjua AA; Zanobetti A; Coull BA; Schwartz J; MacCallum G; Verrier RL; Nearing BD; Canner MJ; Suh H; Stone
PH. (2005). Air pollution and ST-segment depression in elderly subjects. Environ Health Perspect 113: 883-887.
Goldberg MS; Burnett RT; Valois MF; Flegel K; Bailar JC, 3rd; Brook J; Vincent R; Radon K. (2003). Associations between
ambient air pollution and daily mortality among persons with congestive heart failure. Environ Res 91: 8-20.
Goldberg MS; Burnett RT; Yale JF; Valois MF; Brook JR. (2006). Associations between ambient air pollution and daily
mortality among persons with diabetes and cardiovascular disease. Environ Res 100: 255-267.
Gomez-Perales JE; Colvile RN; Fernandez-Bremauntz AA; Gutierrez-Avedoy V; Paramo-Figueroa VH; Blanco-Jimenez S;
Bueno-Lopez E; Bernabe-Cabanillas R; Mandujano F; Hidalgo-Navarro M. (2007). Bus, minibus, metro inter-comparison
of commuters' exposure to air pollution in Mexico City. Atmospheric Environment 41: 890-901.
Gomez-Perales JE; Colvile RN; NieuwenhuijsenMJ; Fernandez-Bremauntz A; Gutierrez-Avedoy VJ; Paramo-Figueroa VH;
Blanco-Jimenez S; Bueno-Lopez E; Mandujano F; Bernabe-Cabanillas R. (2004). Commuters'exposure to PM2. 5, CO,
and benzene in public transport in the metropolitan area of Mexico City. Atmospheric Environment 38: 1219-1229.
Gomot-de Vaufleury A; Kerhoas I. (2000). Effects of cadmium on the reproductive system of the land snail Helix aspersa.
Bulletin of Environmental Contamination and Toxicology 64: 434-442.
Gong H, Jr.; Linn WS; Clark KW; Anderson KR; Sioutas C; Alexis NE; Cascio WE; Devlin RB. (2008). Exposures of healthy
and asthmatic volunteers to concentrated ambient ultrafine particles in Los Angeles. Inhal Toxicol 20: 533-545.
Gong H; Linn WS; Clark KW; Anderson KR; Geller MD; Sioutas C. (2005). Respiratory responses to exposures with fine
particulates and nitrogen dioxide in the elderly with and without COPD. Inhalation Toxicology 17: 123-132.
Gong H; Linn WS; Sioutas C; Terrell SL; Clark KW; Anderson KR; Terrell LL. (2003a). Controlled exposures of healthy and
asthmatic volunteers to concentrated ambient fine particles in Los Angeles. Inhalation Toxicology 15: 305-325.
Gong H; Linn WS; Terrell SL; Anderson KR; Clark KW; Sioutas C; Cascio WE; Alexis N; Devlin RB. (2004a). Exposures of
elderly volunteers with and without chronic obstructive pulmonary disease (COPD) to concentrated ambient fine
particulate pollution. Inhalation Toxicology 16: 731-744.
Gong H; Linn WS; Terrell SL; Clark KW; Geller MD; Anderson KR; Cascio WE; Sioutas C. (2004b). Altered heart-rate
variability in asthmatic and healthy volunteers exposed to concentrated ambient coarse particles. Inhalation Toxicology
16: 335-343.
Gong H; Sioutas C; Linn WS. (2003b). Controlled exposures of healthy and asthmatic volunteers to concentrated ambient
particles in metropolitan Los Angeles. Res Rep Health Eff Inst 118: 1-36; discussion 37-47.
Gong H; Sioutas C; Linn WS; Clark KW; Terrell SL; Terrell LL; Anderson KR; Kim S; Chang MC. (2000). Controlled human
exposures to concentarted ambient fine particles in metropolitan Los Angeles: Methodology and preliminary health-effect
findings. Inhalation Toxicology 12: 107-119.
Gonzalez-Chavez C; Harris PJ; Dodd J; Meharg AA. (2002). Arbuscular mycorrhizal fungi confer enhanced arsenate resistance
on Holcus lanatus. New Phytologist 155: 163-171.
Goodarzi F; Sanei H; Garrett RG; Duncan WF. (2002). Accumulation of trace elements on the surface soil around the Trail
smelter, British Columbia, Canada. Environmental Geology 43: 29-38.
Goodman RM; Yergin BM; Landa JF; Golivanux MH; Sackner MA. (1978). Relationship of smoking history and pulmonary
function tests to tracheal mucous velocity in nonsmokers, young smokers, ex-smokers, and patients with chronic
bronchitis. Am Rev Respir Dis 117: 205-214.
Gordian ME; Choudhury AH. (2003). PM10 and asthma medication in schoolchildren. Archives of environmental health 58: 42-
47.
Gordon T; Gerber H; Fang CP; Chen LC. (1999). A centrifugal particle concentrator for use in inhalation toxicology. Inhal
Toxicol 11: 71-87.
Gordon W; Jackson R. (2000). Nutrient concentrations in fine roots. Ecology 81: 275-280.
Goriaux M; Jourdain B; Temime B; Besombes JL; Marchand N; Albinet A; Leoz-Garziandia E; Wortham H. (2006). Field
comparison of particulate PAH measurements using a low-flow denuder device and conventional sampling systems.
Environ Sci Technol 40: 6398-6404.
December 2008
R-25
DRAFT-DO NOT QUOTE OR CITE
-------
Goss CH; Newsom SA; Schildcrout JS; Sheppard L; Kaufman ID. (2004). Effect of ambient air pollution on pulmonary
exacerbations and lung function in cystic fibrosis. American journal of respiratory and critical care medicine 169: 816-
821.
Gotschi T; Heinrich J; Sunyer J; Kunzli N. (2008). Long-term effects of ambient air pollution on lung function - A review.
Epidemiology 19: 690-701.
Goudie AS; Elaine W; Viles HA. (2002). The roles of salt and fog in weathering: a laboratory simulation of conditions the
northern Atacama Desert, Chile. Catena 48: 255-266.
Gouveia N; Bremner SA; Novaes HM. (2004). Association between ambient air pollution and birth weight in Sao Paulo, Brazil.
Journal of epidemiology and community health 58: 11-17.
Graham B; Guyon P; Taylor PE; Artaxo P; Maenhaut W; Glovsky MM; Flagan RC; Andreae MO. (2003). Organic compounds
present in the natural Amazonian aerosol: Characterization by gas chromatography-mass spectrometry. J Geophys Res
108: 4766.
Grannas AM; Hockaday WC; Hatcher PG; Thompson LG; Mosley-Thompson E. (2006). New revelations on the nature of
organic matter in ice cores. Journal of Geophysical Research 111, D04304, doi:10.1029/2005JD006251.
Grannas AM; Jones AE; Dibb J; Ammann M; Anastasio C; Beine HJ; Bergin M; Bottenheim J; Boxe CS; Carver G; Chen G;
Crawford JH; Domine F; Frey MM; Guzman MI; Heard DE; Helmig D; Hoffmann MR; Honrath RE; Huey LG; Hutterli
M; Jacobi HW; Klan P; Lefer B; McConnell J; Plane J; Sander R; Savarino J; Shepson PB; Simpson WR; Sodeau JR; von
Glasow R; Weller R; Wolff EW; Zhu T. (2007). An overview of snow photochemistry: evidence, mechanisms and
impacts. Atmos Chem Phys 7: 4329-4373.
Grannas AM; Shepson PB; Filley TR. (2004). Photochemistry and nature of organic matter in Arctic and Antarctic snow. Global
Biogeochemical Cycles 18(1), GB1006, doi:10.1029/2003GB002133.
Grantz DA; Garner JHB; Johnson DW. (2003). Ecological effects of particulate matter. Environment International 29: 213-239.
Granville CA; Hanley NM; Mumford JL; DeMarini DM. (2003). Mutation spectra of smoky coal combustion emissions in
Salmonella reflect the TP53 and KRAS mutations in lung tumors from smoky coal-exposed individuals. Mutation
Research-Fundamental and Molecular Mechanisms of Mutagenesis 525: 77-83.
Gray JS. (2002). Biomagnification in marine systems: The perspective of an ecologist. Marine Pollution Bulletin 45: 46-52.
Graydon JA; Louis VLS; Lindberg SE; Hintelmann H; Krabbenhoft DP. (2006). Investigation of mercury exchange between
forest canopy vegetation and the atmosphere using a new dynamic chamber. Environmental Science and Technology 40:
4680-4688.
Greenland P; Kizilbash MA. (2005). Coronary computed tomography in coronary risk assessment. J Cardiopulm Rehabil 25: 3-
10.
Greenspan BJ; Morrow PE; Ferin J. (1988). Effects of Aerosol Exposures to Cadmium Chloride on the Clearance of Titanium
Dioxide from the Lungs of Rats. Experimental lung research 14: 491-499.
Greenstein D; Tiefenthaler L; Bay S. (2004). Toxicity of parking lot runoff after application of simulated rainfall. Archives of
Environmental Contamination and Toxicology 47: 199-206.
Greenwald R; Bergin MH; Weber R; Sullivan A. (2007). Size-resolved, real-time measurement of water-insoluble aerosols in
metropolitan Atlanta during the summer of 2004. Atmospheric Environment 41: 519-531.
Grell GA; Emeis S; Stockwell WR; Schoenemeyer T; Forkel R; Michalakes J; Knoche R; Seidl W. (2000). Application of a
multiscale, coupled MM5/chemistry model to the complex terrain of the VOTALP valley campaign. Atmospheric
Environment 34: 1435-1453.
Grgic B; Finlay WH; Burnell PKP; Heenan AF. (2004). In vitro intersubject and intrasubject deposition measurements in realistic
mouth-throat geometries. J Aerosol Sci 35: 1025-1040.
Griffith DA; Solak ME; Yorty DP. (2005). Is air pollution impacting winter orographic precipitation in Utah? Weather
modification association. Journal of Weather Modification 37: 14-20.
Grigal DF. (2002). Inputs and outputs of mercury from terrestrial watersheds: a review. Environmental Review 10: 1-39.
Grigal DF. (2003). Mercury sequestration in forests and peatlands: A review. J Environ Qual 32: 393-405.
Grigg J; Kulkarni N; Pierse N; Rushton L; O'Callaghan C; Rutman A. (2008). Black-pigmented material in airway macrophages
from healthy children: association with lung function and modeled PM10. Res Rep Health Eff Inst: 1-23; discussion 25-
33.
Groneberg DA; Quarcoo D; Frossard N; Fischer A. (2004). Neurogenic mechanisms in bronchial inflammatory diseases. Allergy
59: 1139-1152.
Grossi CM; Brimblecombe P. (2004). Aesthetics of simulated soiling patterns on architecture. Environmental Science and
Technology 38: 3971-3976.
Grossi CM; Esbert RM; Diaz-Pache F; Alonso FJ. (2003). Soiling of building stones in urban environments. Building and
Environment 38: 147-159.
Grover B; Eatough N; Eatough D; Chow J; Watson J; Ambs J; Meyer M; Hopke P; Al-Horr R; Later D. (2006). Measurement of
Both Nonvolatile and Semi-Volatile Fractions of Fine Particulate Matter in Fresno, CA. Aerosol Science and Technology
40: 811-826.
Grover BD; Hopke PK; Long RW; Wilson WE; Meyer MB; Ambs JL; Kleinman M; Eatough NL; Eatough DJ. (2005).
Measurement of total PM2.5 mass (nonvolatile plus semivolatile) with the Filter Dynamic Measurement System tapered
element oscillating microbalance monitor. Journal of Geophysical Research D: Atmospheres 110: 1-9.
December 2008
R-26
DRAFT-DO NOT QUOTE OR CITE
-------
Grover BD; Kleinman M; Eatough NL; Eatough DJ; Cary RA; Hopke PK; Wilson WE. (2008). Measurement of fine particulate
matter nonvolatile and semi-volatile organic material with the Sunset Laboratory Carbon Aerosol Monitor. J Air Waste
Manag Assoc 58: 72-77.
Grunstein MM; Hazucha M; Sorli MH; Milic-Emili J. (1977). Effect of S02 on control of breathing in anesthetized cats. Journal
of applied physiology 43: 844-851.
GuderianR (Ed.). (1985). Air pollution by photochemical oxidants: formation, transport, control, and effects on plants. New
York, NY: Springer-Verlag.
Guderian R. (1977). Accumulation of pollutants in plant organs. In: Air pollution: phytotoxicity of acidic gases and its
significance in air pollution control. In (Vol. Ecological studies: v. 22, pp. pp. 66-74). Berlin, Germany: Springer-Verlag.
Gulliver J; Briggs DJ. (2004). Personal exposure to particulate air pollution in transport microenvironments. Atmospheric
Environment 38: 1-8.
Gulliver J; Briggs DJ. (2007). Journey-time exposure to particulate air pollution. Atmospheric Environment 41:7195-7207.
Gunnison A; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice. VI. Gene
expression in heart and lung tissue. Inhalation toxicology 17: 225-233.
Gupta P; Christopher SA; Wang J; Gehrig R; Lee Y; Kumar N. (2006). Satellite remote sensing of particulate matter and air
quality assessment over global cities. Atmospheric Environment 40: 5880-5892.
Gurgueira SA; Lawrence J; Coull B; Murthy GG; Gonzalez-Flecha B. (2002). Rapid increases in the steady-state concentration
of reactive oxygen species in the lungs and heart after particulate air pollution inhalation. Environ Health Perspect 110:
749-755.
Gustin MS. (2003). Are mercury emissions from geologic sources significant? A status report. Science of the Total Environment
304: 153-167.
Gustin MS; Engle M; Ericksen J; Lyman S; Stamenkovic J; Xin M. (2006). Mercury exchange between the atmosphere and low
mercury containing substrates. Applied Geochemistry 21: 1913-1923.
Gutierrez-Castillo ME; Roubicek DA; Cebrian-Garcia ME; De Vizcaya-Ruiz A; Sordo-Cedeno M; Ostrosky-Wegman P. (2006).
Effect of chemical composition on the induction of DNA damage by urban airborne particulate matter. Environ Mol
Mutagen 47: 199-211.
Gutierrez-Daban A; Fernandez-Espinosa AJ; Ternero-Rodriguez M; Fernandez-Alvarez F. (2005). Particle-size distribution of
polycyclic aromatic hydrocarbons in urban air in southern Spain. Analytical and Bioanalytical Chemistry 381: 721-736.
Gwynn RC; Burnett RT; Thurston GD. (2000). A time-series analysis of acidic particulate matter and daily mortality and
morbidity in the Buffalo, New York, region. Environ Health Perspect 108: 125-133.
Ha EH; Hong YC; Lee BE; Woo BH; Schwartz J; Christiani DC. (2001). Is air pollution a risk factor for low birth weight in
Seoul? Epidemiology 12: 643-648.
Ha EH; Lee JT; Kim H; Hong YC; Lee BE; Park HS; Christiani DC. (2003). Infant susceptibility of mortality to air pollution in
Seoul, South Korea. Pediatrics 111: 284-290.
Hageman KJ; Simonich SL; Campbell DH; Wilson GR; Landers DH. (2006). Atmospheric deposition of current-use and historic-
use pesticides in snow at national parks in the western United States. Environ Sci Technol 40: 3174-3180.
Hains JC; Chen LWA; Taubman BF; Doddridge BG; Dickerson RR. (2007). A side-by-side comparison of filter-based PM2. 5
measurements ata suburban site: A closure study. Atmospheric Environment 41: 6167-6184.
Haland G; Carlsen KC; Sandvik L; Devulapalli CS; Munthe-Kaas MC; Pettersen M; Carlsen KH. (2006). Reduced lung function
at birth and the risk of asthma at 10 years of age. The New England journal of medicine 355: 1682-1689.
Hall AH; Teschke K; Davies H; Demers P; Marion S. (2002). Exposure levels and determinants of softwood dust exposures in
BC lumber mills. 1981-1997. AIHAJ (Fairfax, Va) 63: 709-714.
Halonen JI; Lanki T; Yli-Tuomi T; Kulmala M; Tiittanen P; Pekkanen J. (2008). Urban air pollution, and asthma and COPD
hospital emergency room visits. Thorax 63: 635-641.
Halsall CJ. (2004). Investigating the occurrence of persistent organic pollutants (POPs) in the Arctic: their atmospheric behaviour
and interaction with the seasonal snow pack. Environmental Pollution 128: 163-175.
Halsall CJ; Barrie LA; Fellin P; Muir DCG; Rovinski FY; Kononov EY. (1997). Spatial and temporal variation of polycyclic
aromatic hydrocarbons in the arctic atmosphere. Environmental Science and Technology 31: 3593-3599.
Hamada K; Suzaki Y; Leme A; Ito T; Miyamoto K; Kobzik L; Kimura H. (2007). Exposure of pregnant mice to an air pollutant
aerosol increases asthma susceptibility in offspring. Journal of toxicology and environmental health Part A 70: 688-695.
Hamade AK; Rabold R; Tankersley CG. (2008). Adverse cardiovascular effects with acute particulate matter and ozone
exposures: interstrain variation in mice. Environ Health Perspect 116: 1033-1039.
Han Y; Cao J; Chow JC; Watson JG; An Z; Jin Z; Fung K; Liu S. (2007). Evaluation of the thermal/optical reflectance method
for discrimination between char-and soot-EC. Chemosphere 69: 569-574.
Hand JL; Kreidenweis SM. (2002). A New Method for Retrieving Particle Refractive Index and Effective Density from Aerosol
Size Distribution Data. Aerosol Science and Technology 36: 1012-1026.
Hand JL; Malm WC. (2006). Review of the IMPROVE Equation for Estimating Ambient Light Extinction Coefficients-Final
Report. Interagency Monitoring of Protective Visual Environments,
Hand JL; Malm WC. (2007). Review of aerosol mass scattering efficiencies from ground-based measurements since 1990.
Journal of Geophysical Research 112.
Hanel B; Law I; Mortensen J. (2003). Maximal rowing has an acute effect on the blood-gas barrier in elite athletes. Journal of
applied physiology 95: 1076-1082.
December 2008
R-27
DRAFT-DO NOT QUOTE OR CITE
-------
Hanigan IC; Johnston FH; Morgan GG. (2008). Vegetation fire smoke, indigenous status and cardio respiratory hospital
admissions in Darwin, Australia, 1996-2005: a time-series study. Environ Health 7: 42.
Hannigan MP; Cass GR; Penman BW; Crespi CL; Lafleur AL; Busby Jr WF; Thilly WG. (1997). Human cell mutagens in Los
Angeles air. ENVIRONMENTAL SCIENCE AND TECHNOLOGY-WASHINGTON DC- 31: 438-447.
Hannigan MP; Cass GR; Penman BW; Crespi CL; Lafleur AL; Busby WF; Thilly WG; Simoneit BRT. (1998). Bioassay-directed
chemical analysis of Los Angeles airborne particulate matter using a human cell mutagenicity assay. Environmental
science & technology 32: 3502-3514.
Hansen C; Neller A; Williams G; Simpson R. (2006). Maternal exposure to low levels of ambient air pollution and preterm birth
in Brisbane, Australia. Bjog 113: 935-941.
Hansen C; Neller A; Williams G; Simpson R. (2007a). Low levels of ambient air pollution during pregnancy and fetal growth
among term neonates in Brisbane, Australia. Environ Res 103: 383-389.
Hansen CS; Sheykhzade M; Moller P; Folkmann JK; Amtorp O; Jonassen T; Loft S. (2007b). Diesel exhaust particles induce
endothelial dysfunction in apoE-/- mice. Toxicology and applied pharmacology 219: 24-32.
Hansen D; Blahout B; Benner D; Popp W. (2008). Environmental sampling of particulate matter and fungal spores during
demolition of a building on a hospital area. J Hosp Infect: epub.
Hansen J; Nazarenko L. (2004). Soot climate forcing via snow and ice albedos. Proc Natl Acad Sci USA 101: 423-428.
Hansen JE; Sato M. (2001). Trends in measured climate forcing agents. Proc Natl Acad Sci USA 98: 778-783.
Hansen K; Draaijers GPJ; Ivens WPMF; Gundersen P; van Leeuwen NFM. (1994). Concentration variations in rain and canopy
throughfall collected sequentially during individual rain events. Atmospheric Environment 28: 3195-3205.
Hao M; Cornier S; Wang M; Lee James J; Nel A. (2003). Diesel exhaust particles exert acute effects on airway inflammation and
function in murine allergen provocation models. Journal of allergy and clinical immunology 112: 905-914.
Hapcioglu B; Issever H; Kocyigit E; Disci R; Vatansever S; Ozdilli K. (2006). The effect of air pollution and meteorological
parameters on chronic obstructive pulmonary disease at an Istanbul Hospital. Indoor and Built Environment 15: 147-153.
Happo MS; Salonen RO; Halinen AI; Jalava PI; Pennanen AS; Kosma VM; Sillanpaa M; Hillamo R; Brunekreef B; Katsouyanni
K; Sunyer J; Hirvonen MR. (2007). Dose and time dependency of inflammatory responses in the mouse lung to urban air
coarse, fine, and ultrafine particles from six European cities. Inhalation toxicology 19: 227-246.
Hargrave BT; Barrie LA; Bidleman TF; Welch HE. (1997). Seasonality in exchange of organochlorines between arctic air and
seawater. Environmental Science and Technology 31: 3258-3266.
Harkema JR; Keeler G; Wagner J; Morishita M; Timm E; Hotchkiss J; Marsik F; Dvonch T; Kaminski N; Barr E. (2004). Effects
of concentrated ambient particles on normal and hypersecretory airways in rats. Res Rep Health Eff Inst: 1-68; discussion
69-79.
Harmens H; Norris DA; Koerber GR; Buse A; Steinnes E; Ruhling A. (2007). Temporal trends in the concentration of arsenic,
chromium, copper, iron, nickel, vanadium and zinc in mosses across Europe between 1990 and 2000. Atmospheric
Environment 41: 6673-6687.
Harner T; Jantunen LMM; Bidleman TF; Barrie LA; Kylin H; Strachan WMJ; Macdonald RW. (2000). Microbial degradation is
a key elimination pathway of hexachlorocyclohexanes from the Arctic Ocean. Geophysical Research Letters 27: 1155-
1158.
Harrison D; Shik Park S; Ondov J; Buckley T; Roul Kim S; Jayanty RKM. (2004). Highly time resolved fine particle nitrate
measurements at the Baltimore Supersite. Atmospheric Environment 38: 5321-5332.
Harrod KS; Jaramillo RJ; Rosenberger CL; Wang S-Z; Berger JA; McDonald JD; Reed MD. (2003). Increased susceptibility to
RSV infection by exposure to inhaled diesel engine emissions. American journal of respiratory cell and molecular
biology 28: 451-463.
Hasegawa S; Hirabayashi M; Kobayashi S; Moriguchi Y; Kondo Y; Tanabe K; Wakamatsu S. (2005). Size Distribution and
Characterization of Ultrafine Particles in Roadside Atmosphere. Journal of Environmental Science and Health, Part A 39:
2671-2690.
Hashimoto AH; Amanuma K; Hiyoshi K; Sugawara Y; Goto S; Yanagisawa R; Takano H; Masumura K; Nohmi T; Aoki Y.
(2007). Mutations in the lungs of gpt delta transgenic mice following inhalation of diesel exhaust. Environ Mol Mutagen
48: 682-693.
Hauck M. (2003). Epiphytic lichen diversity and forest dieback: The role of chemical site factors. Bryologist 106: 257-269.
Hayek T; Oiknine J; Brook JG; Aviram M. (1994). Increased plasma and lipoprotein lipid peroxidation in apo E-deficient mice.
Biochem Biophys Res Commun 201: 1567-1574.
Haynie FH (Ed.). (1986). Environmental factors affecting corrosion of weathering steel (Vol. ACS symposium series 318).
Haynie FH; Lemmons TJ. (1990). Particulate matter soiling of exterior paints at a rural site. Aerosol Science and Technology 13:
356-367.
Hays MD; Geron CD; Linna KJ; Smith ND; Schauer JJ. (2002). Speciation of gas-phase and fine particle emissions from burning
of foliar fuels. EnvironSci Technol 36: 2281-2295.
Hays MD; Lavrich RJ. (2007). Developments in direct thermal extraction gas chromatography-mass spectrometry of fine
aerosols. Trends in Analytical Chemistry 26: 88-102.
Haywood J; Boucher O. (2000). Estimates of the direct and indirect radiative forcing due to tropospheric aerosols: A review. Rev
Geophys 38: 513-543.
Haywood JM; Shine KP. (1997). Multi-spectral calculations of the direct radiative forcing of tropospheric sulphate and soot
aerosols using a column model. Quarterly Journal of the Royal Meteorological Society 123: 1907-1930.
December 2008
R-28
DRAFT-DO NOT QUOTE OR CITE
-------
Healey K; Smith EC; Wild CP; Routledge MN. (2006). The mutagenicity of urban particulate matter in an enzyme free system is
associated with the generation of reactive oxygen species. Mutat Res 602: 1-6.
Hedley AJ; Wong C-M; Thach TQ; Ma S; Lam T-H; Anderson HR. (2002). Cardiorespiratory and all-cause mortality after
restrictions on sulphur content of fuel in Hong Kong: an intervention study. The Lancet 360: 1646-1652.
HEI. Reanalysis of the Harvard Six Cities Study and the American Cancer Society Study of Particulate Air Pollution
and Mortality. 2000. Health Effects Institute.
Heinrich J; Hoelscher B; Frye C; Meyer I; Pitz M; Cyry s J; Wjst M; Neas L; Wichmann HE. (2002). Improved Air Quality in
Reunified Germany and Decreases in Respiratory Symptoms. Epidemiology 13: 394.
Heinrich J; Slama R. (2007). Fine particles, a major threat to children. Int J Hyg Environ Health 210: 617-622.
Heiss G; Sharrett AR; Barnes R; Chambless LE; Szklo M; Alzola C. (1991). Carotid atherosclerosis measured by B-mode
ultrasound in populations: associations with cardiovascular risk factors in the ARIC study. Am J Epidemiol 134: 250-256.
Heistracher T; Hofmann W. (1997). Flow and deposition patterns in successive airway bifurcations. Annals of Occupational
Hygiene 41: 537-542.
Helena P; Franc B; Cvetka RL. (2004). Monitoring of short-term heavy metal deposition by accumulation in epiphytic lichens
(Hypogymnia Physodes (L.) Nyl.). J Atmos Chem 49: 223-230.
Helm PA; Diamond ML; Semkin R; Bidleman TF. (2000). Degradation as a loss mechanism in the fate of a-
hexachlorocyclohexane in Arctic watersheds. Environmental Science and Technology 34: 812-818.
Helm PA; Diamond ML; Semkin R; Strachan WMJ; Teixeira C; Gregor D. (2002). A mass balance model describing multi-year
fate of organochlorine compounds in a high arctic lake. Environmental Science and Technology 36: 996-1003.
Helmisaari H-S; Makkonen K; Olsson M; Viksna A; Malkonen E. (1999). Fine-root growth, mortality and heavy metal
concentrations in limed and fertilized Pinus silvestris (L.) stands in the vicinity of a Cu-Ni smelter in SW Finland. Plant
and Soil 209: 193-200.
Henderson R. (2005). Clean Air Scientific Advisory Committee (CASAC) Review of the EPA Staff Recommendations
Concerning a Potential Thoracic Coarse PM Standard in the Review of the National Ambient Air Quality Standards for
Particulate Matter: Policy Assessment of Scientific and Technical Information (Final PM OAQPS Staff Paper, EPA-
452/R-05-005). September 15, 2005. Available: http://www.epa.gov/sab/panels/casacpmpanel.html.
Henderson R. (2006). Clean Air Scientific Advisory Committee (CASAC) Review of the EPA Staff Recommendations
Concerning a Potential Thoracic Coarse PM Standard in the Review of the National Ambient Air Quality Standards for
Particulate Matter: Policy Assessment of Scientific and Technical Information (Final PM OAQPS Staff Paper, EPA-
452/R-05-005). September 15, 2005. Available: http://www.epa.gov/sab/panels/casacpmpanel.html.
Henneberger A, W. Zareba, A. Ibald-Mulli, R. Ruckerl, J. Cyrys, J.-P. Couderc, B. Mykins, G. Woelke, H.-E. Wichmann, A.
Peters. (2005). Repolarization changes induced by air pollution in ischemic heart disease patients. Environ Health
Perspect 113: 440-446.
Henrotin JB; Besancenot JP; Bejot Y; Giroud M. (2007). Short-term effects of ozone air pollution on ischaemic stroke
occurrence: a case-crossover analysis from a 10-year population-based study in Dijon, France. Occupational and
environmental medicine 64: 439-445.
Henry RC. (1997). History and fundamentals of multivariate air quality receptor models. Chemometrics and Intelligent
Laboratory Systems 37: 37-42.
Hering S; Fine PM; Sioutas C; Jaques PA; Ambs JL; Hogrefe O; Demeijian KL. (2004). Field assessment of the dynamics of
particulate nitrate vaporization using differential TEOM® and automated nitrate monitors. Atmospheric Environment 38:
5183-5192.
Hering SV. (2007). Using Regional Data and Building Leakage to Assess Indoor Concentrations of Particles of Outdoor Origin.
Aerosol Science and Technology 41: 639-654.
Hering SV; Stolzenburg MR; Quant FR; Oberreit DR; Keady PB. (2005). A Laminar-Flow, Water-Based Condensation Particle
Counter (WCPC). Aerosol Science and Technology 39: 659-672.
Hermann M; Wehner B; Bischof O; Han HS; Krinke T; Liu W; Zerrath A; Wiedensohler A. (2007). Particle counting efficiencies
of new TSI condensation particle counters. J Aerosol Sci 38: 674-682.
Hernandez L; Probst A; Probst JL; Ulrich E. (2003). Heavy metal distribution in some French forest soils: Evidence for
atmospheric contamination. Science of the Total Environment 312: 195-219.
Herner J; Ying Q; Aw J; Gao O; Chang D; Kleeman M. (2006). Dominant Mechanisms that Shape the Airborne Particle Size and
Composition Distribution in Central California. Aerosol Science and Technology 40: 827-844.
Herner JD; Aw J; Gao O; Chang DP; Kleeman MX (2005). Size and Composition Distribution of Airborne Particulate Matter in
Northern California: I- Particulate Mass, Carbon, and Water-Soluble Ions. Journal of the Air & Waste Management
Association 55: 30-51.
Herngren L; Goonetilleke A; Ayoko GA. (2006). Analysis of heavy metals in road-deposited sediments. Analytica Chimica Acta
571: 270-278.
Herrera LK; Videla HA. (2004). The importance of atmospheric effects on biodeterioration of cultural heritage constructional
materials. International Biodeterioration and Biodegradation 54: 125-134.
Heywood JM; Shine KP. (1995). The effects of anthropogenic sulfate and soot aerosol on the Clear Sky Planetary Radiation
Budget. Geophysical Research Letters 22: 603-606.
December 2008
R-29
DRAFT-DO NOT QUOTE OR CITE
-------
Hicke JA; Asner GP; Kasischike ES; FrenchNHF; Randerson JT; Collatz GJ; Stocks BJ; Tucker CJ; Los SO; Field CB. (2003).
Postfire response of North American boreal forest net primary productivity analysed with satellite observations. Global
Change Biol 9: 1145-1157.
Hill AB. (1965). The Environment and Disease: Association or Causation? Proc R Soc Med 58: 295-300.
Hiltermann TJ, Stolk J, van der Zee SC, Brunekreef B, de Bruijne CR, Fischer PH, et al. 1998. Asthma severity and
susceptibility to air pollution. EurRespir J 11(3): 686-693.
Hinwood AL; De Klerk N; Rodriguez C; Jacoby P; Runnion T; Rye P; Landau L; Murray F; Feldwick M; Spickett JCEH. (2006).
The relationship between changes in daily air pollution and hospitalizations in Perth, Australia 1992-1998: a case-
crossover study. Int J Environ Health Res 16: 27-46.
Hinzman LD; Fukuda M; Sandberg DV; Chapin III FS; Dash D. (2003). FROSTFIRE: an experimental approach to predicting
the climate feedbacks from the changing boreal fire regime. Journal of Geophysical Research 108 (Dl), 8153,
doi: 10.1029/2001JD000415.
Hiramatsu K; Azuma A; Kudoh S; Desaki M; Takizawa H; Sugawara I. (2003). Inhalation of diesel exhaust for three months
affects major cytokine expression and induces bronchus-associated lymphoid tissue formation in murine lungs.
Experimental lung research 29: 607-622.
Hirose T; Morito K; Kizu R; Toriba A; Hayakawa K; Ogawa S; Inoue S; Muramatsu M; Masamune Y. (2001).
Estrogenic/Antiestrogenic Activities of Benzo [a] pyrene Monohydroxy Derivatives. Journal of health science 47: 552-
558.
Hitzenberger R; Jennings SG; Larson SM; Dillner A; Cachier H; Galambos Z; Roue A; Spain TG. (1999). Intercomparison of
measurement methods for black carbon aerosols. Atmospheric Environment 33: 2823-2833.
Ho JC; Chan KN; Hu WH; Lam WK; Zheng L; Tipoe GL; Sun J; Leung R; Tsang KW. (2001). The effect of aging on nasal
mucociliary clearance, beat frequency, and ultrastructure of respiratory cilia. American journal of respiratory and critical
care medicine 163: 983-988.
Ho KF; Lee SC; Bau KK; Cao JJ; Harrison RME. (2004). Indoor/outdoor relationships of organic carbon (OC) and elemental
carbon (EC) in PM2.5 in roadside environment of Hong Kong. Atmospheric Environment 38: 6327-6335.
Hodzic A; Madronich S; Bohn B; Massie S; Menut L; Wiedinmyer C. (2007). Wildfire particulate matter in Europe during
summer 2003: meso-scale modeling of smoke emissions, transport and radiative effects. Atmos Chem Phys 7: 4043-4064.
Hoek G; Brunekreef B; Goldbohm S; Fischer P; van den Brandt PA. (2002). Association between mortality and indicators of
traffic-related air pollution in the Netherlands: a cohort study. Lancet 360: 1203-1209.
Hoel DG. (1980). Incorporation of background in dose-response models. Fed Proc 39: 73-75.
Hoffmann B; Moebus S; Mohlenkamp S; Stang A; Lehmann N; Dragano N; Schmermund A; Memmesheimer M; Mann K; Erbel
R; Jockel KH. (2007). Residential exposure to traffic is associated with coronary atherosclerosis. Circulation 116: 489-
496.
Hoffmann B; Moebus S; Stang A; Beck EM; Dragano N; Mohlenkamp S; Schmermund A; Memmesheimer M; Mann K; Erbel
R; Jockel KH. (2006). Residence close to high traffic and prevalence of coronary heart disease. European heart journal
27: 2696-2702.
Hoffmann MH; Shi H; Schmitz BL; Schmid FT; Lieberknecht M; Schulze R; Ludwig B; Kroschel U; Jahnke N; Haerer W;
Brambs HJ; Aschoff AJ. (2005). Noninvasive coronary angiography with multislice computed tomography. JAMA 293:
2471-2478.
Hofmann W; Asgharian B. (2003). The Effect of Lung Structure on Mucociliary Clearance and Particle Retention in Human and
Rat Lungs (Vol. 73, pp. 448-456): Soc Toxicology.
Hofmann W; Balashazy I; Heistracher T; Koblinger L. (1996). The Significance of Particle Deposition Patterns in Bronchial
Airway Bifurcations for Extrapolation Modeling. Aerosol Science and Technology 25: 305-327.
Hofmann W; Martonen T; Graham R. (1989). Predicted deposition of nonhygroscopic aerosols in the human lung as a function of
subject age. J Aerosol Med 2: 49-68.
Hogervorst JG; de Kok TM; Briede JJ; Wesseling G; Kleinjans JC; van Schayck CP. (2006). Relationship between radical
generation by urban ambient particulate matter and pulmonary function of school children. J Toxicol Environ Health A
69: 245-262.
Hogrefe C; Hao W; Civerolo K; Ku JY; Sistla G; Gaza RS; Sedefian L; Schere K; Gilliland A; Mathur R. (2007). Daily
Simulation of Ozone and Fine Particulates over New York State: Findings and Challenges. Journal of Applied
Meteorology and Climatology 46: 961-979.
Hogrefe O; Schwab JJ; Drewnick F; Lala GG; Peters S; Demerjian KL; Rhoads K; Felton HD; Rattigan OV; Husain L. (2004).
Semicontinuous PM 2. 5 Sulfate and Nitrate Measurements at an Urban and a Rural Location in New York: PMTACS-
NY Summer 2001 and 2002 Campaigns. Journal of the Air & Waste Management Association 54: 1040-1060.
Holben BN; Eck TF; Slutsker I; Tanre D; Buis JP; Setzer A; Vermote E; Reagan JA; Kaufman YJ; Nakajima T; Lavenu F;
Jankowiak I; Smirnov A. (1998). AERONET: A federated instrument network and data archive for aerosol
characterization. Remote Sensing of Environment 66: 1-16.
Holguin F; Flores S; Ross Z; Cortez M; Molina M; Molina L; Rincon C; Jerrett M; Berhane K; Granados A; Romieu I. (2007).
Traffic-related exposures, airway function, inflammation, and respiratory symptoms in children. American journal of
respiratory and critical care medicine 176: 1236-1242.
Holguin F; Tellez-Rojo MM; Hernandez M; Cortez M; Chow JC; Watson JG; Mannino D; Romieu I. (2003). Air pollution and
heart rate variability among the elderly in Mexico City. Epidemiology 14: 521-527.
December 2008
R-30
DRAFT-DO NOT QUOTE OR CITE
-------
Holland EA; Dentener FJ; Braswell BH; Sulzman JM. (1999). Contemporary and pre-industrial global reactive nitrogen budgets.
Biogeochemistry 46: 7-43.
Hollander M; Hak AE; Koudstaal PJ; Bots ML; Grobbee DE; Hofman A; Witteman JC; Breteler MM. (2003). Comparison
between measures of atherosclerosis and risk of stroke: the Rotterdam Study. Stroke; a journal of cerebral circulation 34:
2367-2372.
Holling CS. (1986). The resilience of terrestrial ecosystems: local surprise and global change. In Clark WC, Munn RE (Eds.),
Sustainable development of the biosphere (pp. pp. 292-317). Cambridge, UK: Cambridge University.
Hong YC; Hwang SS; Kim JH; Lee KH; Lee HJ; Yu SD; Kim DS. (2007). Metals in particulate pollutants affect peak expiratory
flow of schoolchildren. Environ Health Perspect 115: 430-434.
Hopke PK; Ito K; Mar T; Christensen WF; Eatough DJ; Henry RC; Kim E; Laden F; Lall R; Larson TV. (2006). PM source
apportionment and health effects: 1. Intercomparison of source apportionment results. Journal of Exposure Science and
Environmental Epidemiology 16: 275-286.
Hopke PK; Landis MS; Williams RW; Lewis CW; Ramadan Z; Paatero P; Norris GA. (2003). Receptor modeling of ambient and
personal exposure samples: 1998 Baltimore Particulate Matter Epidemiology-Exposure Study. Atmospheric Environment
37: 3289-3302.
Hopke PK; Zhou L; Poirot RL. (2005). Reconciling Trajectory Ensemble Receptor Model Results with Emissions.
Environmental science & technology 39: 7980-7983.
Hornberg C; Maciuleviciute L; Seemayer NH; Kainka E. (1998). Induction of sister chromatid exchanges (SCE) in human
tracheal epithelial cells by the fractions PM-10 and PM-2.5 of airborne particulates. Toxicology Letters 96: 215-220.
Hornberg C; Seemayer L. (1996). Sister chromatid exchanges in rodent tracheal epithelium exposed in vitro to environmental
pollutants. Toxicology Letters 88: 45-53.
Hosker RP, Jr.; Lindberg SE. (1982). Review: atmospheric transport, deposition, and plant assimilation of airborne gases and
particles. Atmospheric Environment 16: 889-910.
Host S; Larrieu S; Pascal L; Blanchard M; Declercq C; Fabre P; Jusot JF; Chardon B; Le Tertre A; Wagner V; Prouvost H;
Lefranc A. (2007). Short-term Associations between Fine and Coarse Particles and Cardiorespiratory Hospitalizations in
Six French Cities. Occupational and environmental medicine 65: 544-551.
Host S; Larrieu S; Pascal L; Blanchard M; Declercq C; Fabre P; Jusot JF; Chardon B; Le Tertre A; Wagner V. (2008). Short-term
associations between fine and coarse particles and hospital admissions for cardiorespiratory diseases in six French cities.
Br Med J 65: 544.
Hougaard KS; Jensen KA; Nordly P; Taxvig C; Vogel U; Saber AT; Wallin H. (2008). Effects of prenatal exposure to diesel
exhaust particles on postnatal development, behavior, genotoxicity and inflammation in mice. Part Fibre Toxicol 5:3.
Houlahan JE; Findlay CS; Schmidt BR; Meyer AH; Kuzmin SL. (2000). Quantitative evidence for global amphibian population
declines. Nature 404: 752-755.
Howe TS; Billings S; Stolzberg RJ. (2004). Sources of polycyclic aromatic hydrocarbons and hexachlorobenzene in spruce
needles of eastern Alaska. Environmental Science and Technology 38: 3294-3298.
Howel D; Moffatt S; Prince H; Bush J; Dunn CE. (2002). Urban Air Quality in North-East England: Exploring the Influences on
Local Views and Perceptions. Risk Analysis 22: 121-130.
HoytDV. (1978). A model for the calculation of solar global insolation. Solar Energy 21: 27-35.
Hsu DJ; Swift DL. (1999). The measurements of human inhalability of ultralarge aerosols in calm air using mannikins. J Aerosol
Sci 30: 1331-1343.
Huang JY; Liao JW; Liu YC; Lu SY; Chou CP; Chan WH; Chen SU; Ueng TH. (2008). Motorcycle Exhaust Induces
Reproductive Toxicity and Testicular Interleukin-6 in Male Rats. Toxicol Sci.
Huang L; Brook JR; Zhang W; Li SM; Graham L; Ernst D; Chivulescu A; Lu G. (2006). Stable isotope measurements of carbon
fractions (OC/EC) in airborne particulate: A new dimension for source characterization and apportionment. Atmospheric
Environment 40: 2690-2705.
Huang SL; Hsu MK; Chan CC. (2003a). Effects of submicrometer particle compositions on cytokine production and lipid
peroxidation of human bronchial epithelial cells. Environmental health perspectives 111: 478-482.
Huang YCT; Ghio AJ; Stonehuerner J; McGee J; Carter JD; Grambow SC; Devlin RB. (2003c). The role of soluble components
in ambient fine particles-induced changes in human lungs and blood. Inhalation Toxicology 15: 327-342.
Huang Y-CT; Soukup J; Harder S; Becker S. (2003b). Mitochondrial oxidant production by a pollutant dust and NO-mediated
apoptosis in human alveolar macrophage. American journal of physiology Cell physiology 284: C24-32.
Huang YL; Batterman S. (2000). Selection and evaluation of air pollution exposure indicators based on geographic areas. Science
of the Total Environment, The 253: 127-144.
Huber SA; Sakkinen P; Conze D; Hardin N; Tracy R. (1999). Interleukin-6 exacerbates early atherosclerosis in mice. Arterioscler
Thromb Vase Biol 19: 2364-2367.
Huchon GJ; Montgomery AB; Lipavsky A; Hoeffel JM; Murray JF. (1987). Respiratory clearance of aerosolized radioactive
solutes of varying molecular weight. J Nucl Med 28: 894-902.
Huchon GJ; Russell JA; Barritault LG; Lipavsky A; Murray JF. (1984). Chronic air-flow limitation does not increase respiratory
epithelial permeability assessed by aerosolized solute, but smoking does. American journal of respiratory and critical care
medicine 130: 457-460.
Huebert BJ; Charlson RJ. (2000). Uncertainties in data on organic aerosols. Tellus B 52: 1249-1255.
December 2008
R-31
DRAFT-DO NOT QUOTE OR CITE
-------
Huebert BJ; Luke WT; Delany AC; Brost RA. (1988). Measurements of concentrations and dry surface fluxes of atmospheric
nitrates in the presence of ammonia. J Geophys Res [Atmos] 93: 7127-7136.
Hughes MK. (1981). Cycling of trace metals in ecosystems. In Lepp NW (Ed.), Effect of heavy metal pollution on plants.
Volume 2: metals in the environment (pp. 95-118). London, United Kingdom: Applied Science Publishers.
Hunaiti AA; Al-Oqlah A; Shannag NM; Abukhalaf IK; Silvestrov NA; Von Deutsch DA; Bayorh MA. (2007). Toward
understanding the influence of soil metals and sulfate content on plant thiols. Journal of Toxicology and Environmental
Health - Part A: Current Issues 70: 559-567.
Hung HH; Halsall CJ; Blanchard P. (2001). Are PCBs in the Canadian Arctic atmosphere declining? Evidence from 5 years of
monitoring. Environmental Science and Technology 35: 1303-1311.
Hung HH; Halsall CJ; Blanchard P; Li HH; Fellin P; Stern G; Rosenberg B. (2002). Temporal trends of organochlorine pesticides
in the Canadian Arctic atmosphere. Environmental Science and Technology 36: 862-868.
Husar RB; TrattDM; Schichtel B; Falke SR; Li F; Jaffe D; Gasso S; Gill T; LaulainenNS; LuF. (2001). Asian dust events of
April 1998. Journal of Geophysical Research 106: 18317-18330.
Hussein T; Johansson C; Karlsson H; Hansson HC. (2008). Factors affecting non-tailpipe aerosol particle emissions from paved
roads: On-road measurements in Stockholm, Sweden. Atmospheric Environment 42: 688-702.
Huynh M; Woodruff TJ; Parker JD; Schoendorf KC. (2006). Relationships between air pollution and preterm birth in California.
Paediatr Perinat Epidemiol 20: 454-461.
Hwang JS; Chan CC. (2002). Effects of air pollution on daily clinic visits for lower respiratory tract illness. Am J Epidemiol 155:
1-10.
Hwang J-S; Nadziejko C; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice.
III. Acute and chronic effects of CAPs on heart rate, heart-rate fluctuation, and body temperature. Inhalation Toxicology
17: 199-207.
Hwang K-W; Lee J-H; Jeong D-Y; Lee C-H; Bhatnagar A; Park J-M; Kim S-H. (2008). Observation of difference in the size
distribution of carbon and major inorganic compounds of atmospheric aerosols after the long-range transport between the
selected days of winter and summer. Atmospheric Environment 42: 1057-1063.
IARC. (2006). IARC monographs on the evaluation of carcinogenic risks to humans: Preamble. . International Agency for
Research on Cancer (IARC). Lyon, France.
Iba MM; Fung J; Chung L; Zhao J; Winnik B; Buckley BT; Chen LC; Zelikoff JT; Kou YR. (2006). Differential inducibility of
rat pulmonary CYP1A1 by cigarette smoke and wood smoke. Mutation research 606: 1-11.
Ibald-Mulli A; Timonen KL; Peters A; Heinrich J; Wolke G; Lanki T; Buzorius G; Kreyling WG; de Hartog J; Hoek G; ten Brink
HM; Pekkanen J. (2004). Effects of particulate air pollution on blood pressure and heart rate in subjects with
cardiovascular disease: a multicenter approach. Environ Health Perspect 112: 369-377.
Ichinose T; Yajima Y; Nagashima M; Takenoshita S; Nagamachi Y; Sagai M. (1997). Lung carcinogenesis and formation of 8-
hydroxy-deoxyguanosine in mice by diesel exhaust particles. Carcinogenesis 18: 185.
ICRP. (1994). Human respiratory tract model for radiological protection. A report of a Task Group of the International
Commission on Radiological Protection. Crit Rev Biotechnol 24: 1-482.
Ilowite JS; Bennett WD; Sheetz MS; Groth ML; Nierman DM. (1989). Permeability of the bronchial mucosa to ? 99 m?Tc-
DTPA in asthma. The American review of respiratory disease 139: 1139-1143.
ILSI. (2000). The relevance of the rat lung response to particle overload for human risk assessment: a workshop consensus report.
ILSI Risk Science Institute Workshop Participants. Inhal Toxicol 12: 1-17.
Imrich A; Ning Y; Lawrence J; Coull B; Gitin E; Knutson M; Kobzik L. (2007). Alveolar macrophage cytokine response to air
pollution particles: oxidant mechanisms. Toxicology and applied pharmacology 218: 256-264.
Ince NH; Dirilgen N; Apikyan IG; Tezcanli G; Ustun B. (1999). Assessment of toxic interactions of heavy metals in binary
mixtures: a statistical approach. Archives of Environmental Contamination and Toxicology 36: 365-372.
Inoue K-I; Takano H; Yanagisawa R; Hirano S; Ichinose T; Shimada A; Yoshikawa T. (2006a). The role of toll-like receptor 4 in
airway inflammation induced by diesel exhaust particles. Archives of toxicology 80: 275-279.
Inoue K-I; Takano H; Yanagisawa R; Hirano S; Sakurai M; Shimada A; Yoshikawa T. (2006b). Effects of airway exposure to
nanoparticles on lung inflammation induced by bacterial endotoxin in mice. Environmental health perspectives 114:
1325-1330.
Inoue K-i; Takano H; Yanagisawa R; Ichinose T; Sadakane K; Yoshino S; Yamaki K; Uchiyama K; Yoshikawa T. (2004).
Components of diesel exhaust particles differentially affect lung expression of cyclooxygenase-2 related to bacterial
endotoxin. Journal of applied toxicology 64: 415-418.
Inoue K-i; Takano H; Yanagisawa R; Sakurai M; Ichinose T; Sadakane K; Yoshikawa T. (2005). Effects of nano particles on
antigen-related airway inflammation in mice. Respiratory research 6.
Inoue K-i; Takano H; Yanagisawa R; Sakurai M; Ueki N; Yoshikawa T. (2007). Effects of diesel exhaust particles on cytokine
production by splenocytes stimulated with lipopolysaccharide. Journal of applied toxicology 27: 95-100.
IOM (2008). Improving the Presumptive Disability Decision-Making Process for Veterans. In Samet James M, Bodurow CC
(Eds.),
IPCC. (2001). Climate Change 2001: The Scientific Basis. Cambridge Univ. Press, Cambridge, UK.
Isaacs KK; Schlesinger RB; Martonen TB. (2006). Three-dimensional computational fluid dynamics simulations of particle
deposition in the tracheobronchial tree. J Aerosol Med 19: 344-352.
December 2008
R-32
DRAFT-DO NOT QUOTE OR CITE
-------
Isakov V; Touma JS; Khlystov A. (2007). A method of assessing air toxics concentrations in urban areas using mobile platform
measurements. JAirWasteManagAssoc 57: 1286-1295.
Ishihara Y; Kagawa J. (2003). Chronic diesel exhaust exposures of rats demonstrate concentration and time-dependent effects on
pulmonary inflammation. Inhalation toxicology 15: 473-492.
Islam T; Gauderman WJ; Berhane K; McConnell R; Avol E; Peters JM; Gilliland FD. (2007). Relationship between air pollution,
lung function and asthma in adolescents. Thorax 62: 957-963.
Ito K. (2003). Associations of Particulate Matter Components with Daily Mortality and Morbidity in Detroit, Michigan. In
Revised Analyses of Time-Series Studies of Air Pollution and Health. Special Report, (pp. 143-156). Boston: Health
Effects Institute.
Ito K; Christensen WF; Eatough DJ; Henry RC; Kim E; Laden F; Lall R; Larson TV; Neas L; Hopke PK; Thurston GD. (2006a).
PM source apportionment and health effects: 2. An investigation of intermethod variability in associations between
source-apportioned fine particle mass and daily mortality in Washington, DC. Journal of Exposure Science and
Environmental Epidemiology 16: 300-310.
Ito K; De Leon S; Thurston GD; Nadas A; Lippmann M. (2005). Monitor-to-monitor temporal correlation of air pollution in the
contiguous US. J ExpoAnalEnvironEpidemiol 15: 172-184.
Ito K; Thurston GD. (1996). Daily PMlO/mortality associations: an investigations of at-risk subpopulations. Journal of exposure
analysis and environmental epidemiology 6: 79-95.
Ito K; Thurston GD; Silverman RA. (2007). Characterization of PM2.5, gaseous pollutants, and meteorological interactions in the
context of time-series health effects models. Journal of Exposure Science and Environmental Epidemiology 17 Suppl 2:
S45-60.
Ito T; Okumura H; Tsukue N; Kobayashi T; Honda K; Sekizawa K. (2006b). Effect of diesel exhaust particles on mRNA
expression of viral and bacterial receptors in rat lung epithelial L2 cells. Toxicology letters 165: 66-70.
Ito T; Suzuki T; Tamura K; Nezu T; Honda K; Kobayashi T. (2008). Examination of mRNA expression in rat hearts and lungs
for analysis of effects of exposure to concentrated ambient particles on cardiovascular function. Toxicology 243: 271-
283.
Jacob D; Avissar R; Bond GC; Gaffin S; Kiehl J; Lean J; Lohmann U; Mann M; Pielke R; Ramanathan V; Russell L. (2005).
Radiative forcing of climate change. Washington, DC: National Academies Press.
Jacob DJ. (1999). Introduction to atmospheric chemistry.
Jacobs SV; Evans GW; Catalano R; Dooley D. (1984). Air pollution and depressive symptomatology: Exploratory analyses of
intervening psychosocial factors. Population & Environment 7: 260-272.
Jacobsen NR; Mrller P; Cohn CA; Loft S; Vogel U; Wallin H. (2008). Diesel exhaust particles are mutagenic in FEl-Muta™
Mouse lung epithelial cells. Mutation Research-Fundamental and Molecular Mechanisms of Mutagenesis.
Jacobson MZ. (2000). A physically based treatment of elemental carbon optics: implications for global direct forcing of aerosols.
Geophysical Research Letters 27: 217-220.
Jacobson MZ. (2001). GATOR-GCMM: a global through urban scale air pollution and weather forecast model: 1. Model design
and treatment of subgrid soil, vegetation, roads, rooftops, water, sea ice, and snow. Journal of Geophysical Research 106:
5385-5402.
Jacobson MZ. (2002a). Atmospheric Pollution: History, Science, and Regulation: Cambridge University Press.
Jacobson MZ. (2002b). Control of fossil-fuel particulate black carbon and organic matter, possibly the most effective method of
slowing global warming. Journal of Geophysical Research 107: 4410.
Jacobson MZ. (2003a). Development of mixed-phase clouds from multiple aerosol size distributions and the effect of the clouds
on aerosol removal. Journal of Geophysical Research 108 D8, 4245, doi:10.1029/2002JD002691.
Jacobson MZ. (2003b). Reply to comment by D. P. Chock et al. on "Control of fossil-fuel particulate black carbon and organic
matter, possibly the most effective method of slowing global warming". Journal of Geophysical Research 108 D24, 4770,
doi: 10.1029/2003JD003707.
Jacobson MZ. (2003c). Reply to comment by J. E. Penner on "Control of fossil-fuel particulate black carbon and organic matter,
possibly the most effective method of slowing global warming". Journal of Geophysical Research 108 (D24), 4772,
doi:10.1029/2003JD003403.
Jacobson MZ. (2003d). Reply to comment by J. Feichter et al. on "Control of fossil-fuel particulate black carbon and organic
matter, possibly the most effective method of slowing global warming". Journal of Geophysical Research 108 D24, 4768,
doi: 10.1029/2002JD003299,.
Jacobson MZ. (2004). Climate response of fossil fuel and biofuel soot, accounting for soot's feedback to snow and sea ice albedo
and emissivity. Journal of Geophysical Research 109, D21201, doi:10.1029/2004JD004945.
Jacobson MZ. (2006). Effects of externally-through-internally-mixed soot inclusions within clouds and precipitation on global
climate. JPhysChemA 110: 6860-6873.
Jacobson MZ; Kaufman YJ. (2006). Wind reduction by aerosol particles. Geophysical Research Letters 33: L24814,
doi:24810.21029/22006GL027838.
Jaffe DH; Singer ME; Rimm AA. (2003). Air pollution and emergency department visits for asthma among Ohio Medicaid
recipients, 1991-1996. Environ Res 91: 21-28.
Jalaludin B; Mannes T; Morgan G; Lincoln D; Sheppeard V; Corbett S. (2007). Impact of ambient air pollution on gestational
age is modified by season in Sydney, Australia. Environ Health 6: 16.
December 2008
R-33
DRAFT-DO NOT QUOTE OR CITE
-------
Jalaludin B; Morgan G; Lincoln D; Sheppeard V; Simpson R; Corbett S. (2006). Associations between ambient air pollution and
daily emergency department attendances for cardiovascular disease in the elderly (65+ years), Sydney, Australia. Journal
of Exposure Science and Environmental Epidemiology 16: 225-237.
Jalaludin BB; O'Toole BI; Leeder SR. (2004). Acute effects of urban ambient air pollution on respiratory symptoms, asthma
medication use, and doctor visits for asthma in a cohort of Australian children. Environ Res 95: 32-42.
Jalava P; Salonen RO; Halinen AI; Sillanpaa M; Sandell E; Hirvonen MR. (2005). Effects of sample preparation on chemistry,
cytotoxicity, and inflammatory responses induced by air particulate matter. Inhalation toxicology 17: 107-117.
Jalava PI; Salonen RO; Halinen AI; Penttinen P; Pennanen AS; Sillanpaa M; Sandell E; Hillamo R; Hirvonen M-R. (2006). In
vitro inflammatory and cytotoxic effects of size-segregated particulate samples collected during long-range transport of
wildfire smoke to Helsinki. Toxicology and applied pharmacology 215: 341-353.
Jalava PI; Salonen RO; Pennanen AS; Happo MS; Penttinen P; Halinen AI; Sillanpaa M; Hillamo R; Hirvonen MR. (2008).
Effects of solubility of urban air fine and coarse particles on cytotoxic and inflammatory responses in RAW 264.7
macrophage cell line. Toxicol Appl Pharmacol 229: 146-160.
Jalava PI; Salonen RO; Pennanen AS; Sillanpaa M; Halinen AI; Happo MS; Hillamo R; Brunekreef B; Katsouyanni K; Sunyer J;
Hirvonen M-R. (2007). Heterogeneities in inflammatory and cytotoxic responses of RAW 264.7 macrophage cell line to
urban air coarse, fine, and ultrafine particles from six European sampling campaigns. Inhalation toxicology 19: 213-225.
James DS; Lambert WE; Mermier CM; Stidley CA; Chick TW; Samet JM. (1997). Oronasal distribution of ventilation at
different ages. Archives of environmental health 52: 118-123.
James SM; Little EE; Semlitsch RD. (2004). Effects of multiple routes of cadmium exposure on the hibernation success of the
American toad (Bufo americanus). Archives of Environmental Contamination and Toxicology 46: 518-527.
Janes H; Dominici F; Zeger SL. (2007). Trends in air pollution and mortality: an approach to the assessment of unmeasured
confounding. Epidemiology 18: 416-423.
Janes H; Sheppard L; Lumley T. (2005). Case-crossover analyses of air pollution data: reference selection strategies and their
implications for bias. Epidemiology 16: 717-726.
Jansen KL; Larson TV; Koenig JQ; Mar TF; Fields C; Stewart J; Lippmann M. (2005). Associations between health effects and
particulate matter and black carbon in subjects with respiratory disease. Environmental Health Perspectives 113: 1741-
1746.
Janssen NA; Brunekreef B; van Vliet P; Aarts F; Meliefste K; Harssema H; Fischer P. (2003). The relationship between air
pollution from heavy traffic and allergic sensitization, bronchial hyperresponsiveness, and respiratory symptoms in Dutch
schoolchildren. Environ HealthPerspect 111: 1512-1518.
Janssen NA; Schwartz J; Zanobetti A; Suh HH. (2002). Air conditioning and source-specific particles as modifiers of the effect
of PM(10) on hospital admissions for heart and lung disease. Environ Health Perspect 110: 43-49.
Jantunen LMM; Bidleman TF. (1996). Air-water gas exchange of hexachlorocyclohexanes (HCHs) and the enantiomers of a-
HCH in Arctic regions. Journal of Geophysical Research 101: 28837-28846.
Jantunen LMM; Bidleman TF. (1998). Organochlorine pesticides and enantiomers of chiral pesticides in Arctic Ocean water.
Archives of Environmental Contamination and Toxicology 35.
Jaques PA; Ambs JL; Grant WL; Sioutas C. (2004). Field evaluation of the differential TEOM monitor for continuous PM 2.5
mass concentrations. Aerosol Science and Technology 38: 49-59.
Jarabek AM; Asgharian B; Miller FJ. (2005). Dosimetric adjustments for interspecies extrapolation of inhaled poorly soluble
particles (PSP). Inhalation Toxicology 17: 317-334.
Jaspers I; Ciencewicki JM; Zhang W; Brighton LE; Carson JL; Beck MA; Madden MC. (2005). Diesel exhaust enhances
influenza virus infections in respiratory epithelial cells. Toxicol Sci 85: 990-1002.
Jayanty R. (2003). Overview of PM2 5 chemical speciation network program. Paper presented at the Ameican Association for
Aerosol Research.
Jeffers DE. (2005). Relative magnitudes of the effects of electrostatic image and thermophoretic forces on particles in the
respiratory tract (Vol. 113, pp. 189-194): NTP.
Jerrett M; Burnett RT; Ma R; Pope CA, 3rd; Krewski D; Newbold KB; Thurston G; Shi Y; Finkelstein N; Calle EE; Thun MJ.
(2005b). Spatial analysis of air pollution and mortality in Los Angeles. Epidemiology 16: 727-736.
Jerrett M; Buzzelli M; Burnett RT; DeLuca PF. (2005a). Particulate air pollution, social confounders, and mortality in small areas
of an industrial city. Soc Sci Med 60: 2845-2863.
Jiang J; Oberdrster G; Elder A; Gelein R; Mercer P; Biswas P. (2008). Does nanoparticle activity depend upon size and crystal
phase? Nanotoxicology 2: 33-42.
Jiao XC; Xu FL; Dawson R; Chen SH; Tao S. (2007). Adsorption and absorption of polycyclic aromatic hydrocarbons to rice
roots. Environmental Pollution 148: 230-235.
Jimenez JL; Jayne JT; Shi Q; Kolb CE; Worsnop DR; Yourshaw I; Seinfeld JH; Flagan RC; Zhang X; Smith KA. (2003).
Ambient aerosol sampling using the Aerodyne Aerosol Mass Spectrometer. J Geophys Res 108: 8425.
Jirak IL; Cotton WR. (2006). Effect of air pollution on precipitation along the Front Range of the Rocky Mountains. Journal of
Applied Meteorology 45: 236-245.
Johnson D; Hale B. (2008). Fine root decomposition and cycling of Cu, Ni, Pb, and Zn at forest sites near smelters in Sudbury,
ON, and Rouyn-Noranda, QU, Canada. Human and Ecological Risk Assessment 14: 41-53.
December 2008
R-34
DRAFT-DO NOT QUOTE OR CITE
-------
Johnson MS; Franke LS; Lee RB; Holladay SD. (1999). Bioaccumulation of 2,4,6-trinitrotoluene and polychlorinated biphenyls
through two routes of exposure in a terrestrial amphibian: Is the dermal route significant? Environmental Toxicology and
Chemistry 18: 873-878.
Johnston FH, Webby RJ, Pilotto LS, Bailie RS, Parry DL, Halpin SJ. 2006. Vegetation fires, particulate air pollution
and asthma: a panel study in the Australian monsoon tropics. Int J Environ Health Res 16(6): 391-404.
Johnston FH; Bailie RS; Pilotto LS; Hanigan IC. (2007). Ambient biomass smoke and cardio-respiratory hospital admissions in
Darwin, Australia. BMC public health 7: 240.
Jones AM; Harrison RM. (2006). Assessment of natural components of PM10 at UK urban and rural sites. Atmospheric
Environment 40: 7733-7741.
Jones GS; Jones A; Roberts DL; Stott PA; Williams KD. (2005). Sensitivity of global-scale climate change attribution results to
inclusion of fossil fuel black carbon aerosol. Geophysical Research Letters 32, L14701, doi:10.1029/2005GL023370.
Jones JG; Minty BD; Lawler P; Hulands G; Crawley JC; Veall N. (1980). Increased alveolar epithelial permeability in cigarette
smokers. Lancet 1: 66-68.
Jones JG; Minty BD; Royston D; Royston JP. (1983). Carboxyhaemoglobin and pulmonary epithelial permeability in man.
Thorax 38: 129-133.
Joseph PM. (2008). Can fine particulate matter explain the paradoxical ozone associations? Environment International 34: 1185-
1191.
Joynt J; Bischoff M; Turco R; Konopka A; Nakatsu CH. (2006). Microbial community analysis of soils contaminated with lead,
chromium and petroleum hydrocarbons. Microbial Ecology 51: 209-219.
Just J; Segala C; Sahraoui F; Priol G; Grimfeld A; Neukirch F. (2002). Short-term health effects of particulate and photochemical
air pollution in asthmatic children. Eur Respir J 20: 899-906.
Kaan PM; Hegele RG. (2003). Interaction between respiratory syncytial virus and particulate matter in guinea pig alveolar
macrophages. American journal of respiratory cell and molecular biology 28: 697-704.
Kabir Z; Bennett K; Clancy L. (2007). Lung cancer and urban air-pollution in Dublin: a temporal association? Ir Med J 100: 367-
369.
Kalberer M; Paulsen D; Sax M; Steinbacher M; Dommen J; Prevot ASH; Fisseha R; Weingartner E; Frankevich V; Zenobi R.
(2004). Identification of Polymers as Major Components of Atmospheric Organic Aerosols (Vol. 303, pp. 1659-1662):
American Association for the Advancement of Science.
Kamh GME. (2005). The impact of landslides and salt weathering on Roman structures at high latitudes - Conway Castle, Great
Britain: A case study. Environmental Geology 48: 238-254.
Kan H; London SJ; Chen G; Zhang Y; Song G; Zhao N; Jiang L; Chen B. (2008). Season, sex, age, and education as modifiers of
the effects of outdoor air pollution on daily mortality in Shanghai, China: The Public Health and Air Pollution in Asia
(PAPA) Study. Environ Health Perspect 116: 1183-1188.
Kanakidou M; Seinfeld JH; Pandis SN; Barnes I; Dentener FJ; Facchini MC; Van Dingenen R; Ervens B; Nenes A; Nielsen CJ;
Swietlicki E; Putaud JP; Balkanski Y; Fuzzi S; Horth J; Moortgat GK; Winterhalter R; Myhre CEL; Tsigaridis K; Vignati
E; Stephanou EG; Wilson J. (2005). Organic aerosol and global climate modelling: a review. Atmos Chem Phys 5: 1053-
1123.
Kandeler E; Kampichler C; Horak O. (1996). Influence of heavy metals on the functional diversity of soil microbial
communities. Biology and Fertility of Soils 23: 299-306.
Kannel WB; Abbott RD; Savage DD; McNamara PM. (1983). Coronary heart disease and atrial fibrillation: the Framingham
Study. Am Heart J 106: 389-396.
Kapp N; Kreyling W; Schulz H; ImHof V; GehrP; Semmler M; Geiser M. (2004). Electron energy loss spectroscopy for
analysis of inhaled ultrafine particles in rat lungs. MicroscRes Tech 63: 298-305.
Kappos AD; Bruckmann P; Eikmann T; Englert N; Heinrich U; Hoeppe P; Koch E; Krause GHM; Kreyling WG; Rauchfuss K;
Rombout P; Schulz-Klemp V; Thiel WR; Wichmann H. (2004). Health effects of particles in ambient air. International
Journal of Hygiene and Environmental Health 207: 399-407.
Karar K; Gupta AK; Kumar A; Biswas AK. (2006). Characterization and identification of the sources of chromium, zinc, lead,
cadmium, nickel, manganese and iron in PM10 particulates at the two sites of Kolkata, India. Environmental Monitoring
and Assessment 120: 347-360.
Karoly ED; Li Z; Dailey LA; Hyseni X; Huang YCT. (2007). Up-regulation of Tissue Factor in Human Pulmonary Artery
Endothelial Cells after Ultrafine Particle Exposure. Environmental health perspectives 115: 535.
Karr C; Lumley T; Schreuder A; Davis R; Larson T; Ritz B; Kaufman J. (2007). Effects of subchronic and chronic exposure to
ambient air pollutants on infant bronchiolitis. Am J Epidemiol 165: 553-560.
Karthikeyan VJ; Lip GYH. (2007). Peripheral artery disease and hypertension: the relation between ankle-brachial index and
mortality. J Hum Hypertens 21: 762-765.
Kastner-Klein P; Plate EJ. (1999). Wind-tunnel study of concentration fields in street canyons. Atmospheric Environment 33:
3973-3979.
Kato A; Kagawa J. (2003). Morphological effects in rat lungs exposed to urban roadside air. Inhalation toxicology 15: 799-818.
Katsouyanni K; Touloumi G; Samoli E; Petasakis Y; Analitis A; Le Tertre A; Rossi G; Zmirou D; Ballester F; Boumghar A.
(2003). Sensitivity analysis of various models of short-term effects of ambient particles on total mortality in 29 cities in
APHEA2. Health effects institute: 157-164.
December 2008
R-35
DRAFT-DO NOT QUOTE OR CITE
-------
Kauffmann F. (2004). Post-genome respiratory epidemiology: a multidisciplinary challenge. European Respiratory Journal 24:
471-480.
Kaufman YJ, et al. (2005). Aerosol anthropogenic component estimated from satellite data. Geophysical Research Letters 32,
L17804, doi:10.1029/2005GL023125.
Kaufman YJ. (2006). Satellite observations of natural and anthropogenic aerosol effects on clouds and climate. Space Science
Reviews 125: 139-147.
Kaul N; Forman HJ. (1996). Activation of NF?B by the respiratory burst of macrophages. Free Radical Biology and Medicine 21:
401-405.
Kaur S; Nieuwenhuijsen M; Colvile R. (2005b). Personal exposure of street canyon intersection users to PM2.5, ultrafme particle
counts and carbon monoxide in Central London, UK. Atmospheric Environment 39: 3629-3641.
Kaur S; Nieuwenhuijsen MJ; Colvile RN. (2005a). Pedestrian exposure to air pollution along a major road in Central London,
UK. Atmospheric Environment 39: 7307-7320.
Kavouras IG; Etyemezian V; Xu J; DuBois DW; Green M; Pitchford M. (2007). Assessment of the local windblown component
of dust in the western United States. Journal of Geophysical Research D Atmospheres 112.
Kaya E; Fidan F; Unlu M; Sezer M; Tetik L; Acar M. (2006). Evaluation of alveolar clearance by Tc-99m DTPA radioaerosol
inhalation scintigraphy in welders. ANNALS OF NUCLEAR MEDICINE 20: 503.
Kaya M; Salan A; Tabakoglu E; Aydogdu N; Berkarda S. (2003). The bronchoalveolar epithelial permeability in house painters
as determined by Tc-99m DTPA aerosol scintigraphy. ANNALS OF NUCLEAR MEDICINE 17: 305-308.
Keatinge WR; Donaldson GC. (2006). Heat acclimatization and sunshine cause false indications of mortality due to ozone.
Environ Res 100: 387-393.
Kehrl HR; Vincent LM; Kowalsky RJ; Horstman DH; O'Neil JJ; McCartney WH; Bromberg PA. (1987). Ozone exposure
increases respiratory epithelial permeability in humans. Am Rev Respir Dis 135: 1124-1128.
Keller A; Siegmann HC. (2001). The role of condensation and coagulation in aerosol monitoring. Journal of exposure analysis
and environmental epidemiology 11: 441-448.
Kelly JT; Asgharian B; Wong BA. (2005). Inertial Particle Deposition in a Monkey Nasal Mold Compared with that in Human
Nasal Replicas. Inhalation Toxicology 17: 823-830.
Kenny LC; Merrifield T; Mark D; Gussman R; Thorpe A. (2004). The Ddevelopment and Designation Testing of a New USEPA-
Approved Fine Particle Inlet: A Study of the USEPA Designation Process. Aerosol Science and Technology 38: 15-22.
Kettunen J; Lanki T; Tiittanen P; Aalto PP; Koskentalo T; Kulmala M; Salomaa V; Pekkanen J. (2007). Associations of fine and
ultrafme particulate air pollution with stroke mortality in an area of low air pollution levels. Stroke; a journal of cerebral
circulation 38: 918-922.
Ketzel M; Wahlin P; Berkowicz R; Palmgren F. (2003). Particle and trace gas emission factors under urban driving conditions in
Copenhagen based on street and roof-level observations. Atmospheric Environment 37: 2735-2749.
Khattar RS; Swales JD; Dore C; Senior R; Lahiri A. (2001). Effect of aging on the prognostic significance of ambulatory
systolic, diastolic, and pulse pressure in essential hypertension. Circulation 104: 783-789.
Khlystov A; Stanier C; Pandis SN. (2004). An Algorithm for Combining Electrical Mobility and Aerodynamic Size Distributions
Data when Measuring Ambient Aerosol. Aerosol Science and Technology 38: 229-238.
Khlystov A; Stanier CO; Takahama S; Pandis SN. (2005). Water content of ambient aerosol during the Pittsburgh Air Quality
Study. J Geophys Res 110: D07S10.
Khoury Z; Schwartz R; Gottlieb S; Chenzbraun A; Stern S; Keren A. (1997). Relation of coronary artery disease to
atherosclerotic disease in the aorta, carotid, and femoral arteries evaluated by ultrasound. Am J Cardiol 80: 1429-1433.
Kidwell CB; Ondov JM. (2004). Elemental Analysis of Sub-Hourly Ambient Aerosol Collections. Aerosol Science and
Technology 38: 205-218.
Kiikkila O. (2003). Heavy-metal pollution and remediation of forest soil around the Harjavalta Cu-Ni smelter, in SW Finland.
SilvaFennica 37: 399-415.
Kim CG; Bell JNB; Power SA. (2003). Effects of soil cadmium on Pinus sylvestris L. seedlings. Plant and Soil 257: 443-449.
Kim CS. (2000). Respiratory dose of inhaled ultrafme particles in healthy adults. Philosophical Transactions of the Royal Society
A: Mathematical, Physical and Engineering Sciences 358: 2693-2705.
Kim CS; Hu SC. (1998). Regional deposition of inhaled particles in human lungs: comparison between men and women (Vol. 84,
pp. 1834-1844): Am Physiological Soc.
Kim CS; Iglesias AJ. (1989). Deposition of inhaled particles in bifurcating airway models: I. Inspiratory deposition. J Aerosol
Med 2: 1-14.
Kim CS; Iglesias AJ; Garcia L. (1989). Deposition of inhaled particles in bifurcating airway models: II. Expiratory deposition. J
Aerosol Med 2: 15-27.
Kim D; Sass-Kortsak A; Purdham JT; Dales RE; Brook JR. (2005a). Sources of personal exposure to fine particles in Toronto,
Ontario, Canada. J Air Waste ManagAssoc 55: 1134-1146.
Kim E; Hopke PK; Pinto JP; Wilson WE. (2005b). Spatial variability of fine particle mass, components, and source contributions
during the Regional Air Pollution Study in St. Louis. Environmental Science and Technology 39: 4172-4179.
Kim JJ; Smorodinsky S; Lipsett M; Singer BC; Hodgson AT; Ostro B. (2004). Traffic-related air pollution near busy roads: the
East Bay Children's Respiratory Health Study. American journal of respiratory and critical care medicine 170: 520-526.
Kim OJ; Ha EH; Kim BM; Park HS; Jung WJ; Lee BE; Suh YJ; Kim YJ; Lee JT; Kim H; Hong YC. (2007a). PM10 and
pregnancy outcomes: a hospital-based cohort study of pregnant women in Seoul. J Occup Environ Med 49: 1394-1402.
December 2008
R-36
DRAFT-DO NOT QUOTE OR CITE
-------
Kim SY; O'Neill MS; Lee JT; Cho Y; Kim J; Kim H. (2007b). Air pollution, socioeconomic position, and emergency hospital
visits for asthma in Seoul, Korea. Int Arch Occup Environ Health 80: 701-710.
Kim Y; Hatsushika H; Muskett RR; Yamazaki K. (2005c). Possible effect of boreal wildfire soot on Arctic sea ice and Alaska
glaciers. Atmospheric Environment 39: 3513-3520.
Kimbell JS. (2006). Nasal Dosimetry of Inhaled Gases and Particles: Where Do Inhaled Agents Go in the Nose? Toxicologic
pathology 34: 270.
Kinne S, et al. (2006). An AeroCom initial assessment: optical properties in aerosol component modules of global models.
Atmospheric Chemistry and Physics 6: 1815-1834.
Kinney PL; Aggarwal M; Northridge ME; Janssen NA; Shepard P. (2000). Airborne Concentrations of PM~ 2~.~ 5 and Diesel
Exhaust Particles on Harlem Sidewalks: A Community-Based Pilot Study. Environmental health perspectives 108: 213-
218.
Kinsey JS; Linna K; King FG; Logan R; Dong Y; Thompson GJ; Clark NN; Mitchell WA; Squier WC. (2006). Evaluation of
methods for the determination of diesel-generated fine particulate matter: Physical characterization results. J Aerosol Sci
37: 63-87.
Kirkbride MP; Dugmore JA. (2003). Glaciological response to distal tephra fallout from the 1947 eruption of Hekla, south
Iceland. Journal of Glaciology 49: 420-428.
Kiss G; Varga B; Galambos I; Ganszky I. (2002). Characterization of water-soluble organic matter isolated from atmospheric
fine aerosol. Journal of Geophysical Research (Atmospheres 107.
Kittelson DB. (1998). Engines and nanoparticles a review. J Aerosol Sci 29: 575-588.
Kittelson DB; Schauer JJ; Lawson DR; Watts WF; Johnson JP. (2006a). On-road and laboratory evaluation of combustion
aerosols-Part 2:. Summary of spark ignition engine results. J Aerosol Sci 37: 931-949.
Kittelson DB; Watts WF; Johnson JP. (2006b). On-road and laboratory evaluation of combustion aerosols—Parti: Summary of
diesel engine results. J Aerosol Sci 37: 913-930.
Kizu R; Okamura K; Toriba A; Mizokami A; Burnstein KL; Klinge CM; Hayakawa K. (2003). Antiandrogenic activities of
diesel exhaust particle extracts in PC3/AR human prostate carcinoma cells. Toxicological sciences 76: 299-309.
Kleeberger SR; Ohtsuka Y. (2005). Gene-particulate matter-health interactions. Toxicology and applied pharmacology 207: 276-
281.
Kleeman MJ; Hughes LS; Allen JO; Cass GR. (1999). Source contributions to the size and composition distribution of
atmospheric particles: Southern California in September 1996. Environmental science & technology 33: 4331-4341.
Kleindienst TE; Edney EO; Lewandowski M; Offenberg JH; Jaoui M. (2006). Secondary Organic Carbon and Aerosol Yields
from the Irradiations of Isoprene and-Pinene in the Presence of NO x and SO 2. Environ Sci Technol 40: 3807-3812.
Kleinman MT; Hamade A; Meacher D; Oldham M; Sioutas C; Chakrabarti B; Stram D; Froines JR; Cho AK. (2005). Inhalation
of concentrated ambient particulate matter near a heavily trafficked road stimulates antigen-induced airway responses in
mice. J Air Waste Manag Assoc 55: 1277-1288.
Kleinman MT; Hyde DM; Bufalino C; Basbaum C; Bhalla DK; Mautz WJ. (2003b). Toxicity of chemical components of fine
particles inhaled by aged rats: effects of concentration. Journal of the Air & Waste Management Association 53: 1080-
1087.
Kleinman MT; Sioutas C; Chang MC; Boere AJF; Cassee FR. (2003a). Ambient fine and coarse particle suppression of alveolar
macrophage functions. Toxicology letters 137: 151-158.
Kleinman MT; Sioutas C; Froines JR; Fanning E; Hamade A; Mendez L; Meacher D; Oldham M. (2007). Inhalation of
concentrated ambient particulate matter near a heavily trafficked road stimulates antigen-induced airway responses in
mice. Inhalation toxicology 19 Suppl 1: 117-126.
Klein-Patel ME; Diamond G; Boniotto M; Saad S; Ryan LK. (2006). Inhibition of beta-defensin gene expression in airway
epithelial cells by low doses of residual oil fly ash is mediated by vanadium. Toxicological sciences 92: 115-125.
Klemm RJ; Lipfert FW; Wyzga RE; Gust C. (2004). Daily mortality and air pollution in Atlanta: two years of data from ARIES.
Inhal Toxicol 16 Suppl 1: 131-141.
Klepeis NE; Nelson WC; Ott WR; Robinson JP; Tsang AM; Switzer P; Behar JV; Hern SC; Engelmann WH. (2001). The
National Human Activity Pattern Survey (NHAPS): a resource for assessing exposure to environmental pollutants.
Journal of exposure analysis and environmental epidemiology 11: 231-252.
Klouda GA; Filliben JJ; Parish HJ; Chow JC; Watson JG; Cary RA. (2005). Reference Material 8785: Air Particulate Matter on
Filter Media. Aerosol Science and Technology 39: 173-183.
Knight EL; Curhan GC. (2003). Albuminuria: moving beyond traditional microalbuminuria cut-points. Curr Opin Nephrol
Hypertens 12: 283-284.
Knobel HH; Chen CJ; Liang KY. (1995). Sudden infant death syndrome in relation to weather and optimetrically measured air
pollution in Taiwan. Pediatrics 96: 1106-1110.
Ko FW; Tam W; Wong TW; Chan DP; Tung AH; Lai CK; Hui DS. (2007b). Temporal relationship between air pollutants and
hospital admissions for chronic obstructive pulmonary disease in Hong Kong. Thorax 62: 780-785.
Ko FW; Tam W; Wong TW; Lai CK; Wong GW; Leung TF; Ng SS; Hui DS. (2007a). Effects of air pollution on asthma
hospitalization rates in different age groups in Hong Kong. Clin Exp Allergy 37: 1312-1319.
Kobayashi H; Matsunaga T; Hoyano A; Aoki M; Komori D; Boonyaway S. (2004). Satellite estimation of phyosynthetically
active radiation in Southeast Asia: impacts of smoke and cloud cover. Journal of Geophysical Research 109.
December 2008
R-37
DRAFT-DO NOT QUOTE OR CITE
-------
Kodavanti UP; Schladweiler MC; Gilmour PS; Wallenborn JG; Mandavilli BS; Ledbetter AD; CMstiani DC; Runge MS; Karoly
ED; Costa DL; Peddada S; Jaskot R; Richards JH; Thomas R; Madamanchi NR; Nyska A. (2008). The role of particulate
matter-associated zinc in cardiac injury in rats. Environmental health perspectives 116: 13-20.
Kodavanti UP; Schladweiler MC; Ledbetter AD; Hauser R; Christiani DC; McGee J; Richards JR; Costa DL. (2002). Temporal
association between pulmonary and systemic effects of particulate matter in healthy and cardiovascular compromised
rats. Journal of toxicology and environmental health part A 65: 1545-1569.
Kodavanti UP; Schladweiler MC; Ledbetter AD; McGee JK; Walsh L; Gilmour PS; Highfill JW; Davies D; Pinkerton KE;
Richards JH; Crissman K; Andrews D; Costa DL. (2005). Consistent pulmonary and systemic responses from inhalation
of fine concentrated ambient particles: roles of rat strains used and physicochemical properties. Environmental health
perspectives 113: 1561-1568.
Koenig JQ; Jansen K; Mar TF; Lumley T; Kaufman J; Trenga CA; Sullivan J; Liu LJ; Shapiro GG; Larson TV. (2003).
Measurement of offline exhaled nitric oxide in a study of community exposure to air pollution. Environ Health Perspect
111: 1625-1629.
Koenig JQ; Mar TF; Allen RW; Jansen K; Lumley T; Sullivan JH; Trenga CA; Larson T; Liu LJ. (2005). Pulmonary effects of
indoor- and outdoor-generated particles in children with asthma. Environ Health Perspect 113: 499-503.
Koike E; Hirano S; Furuyama A; Kobayashi T. (2004). cDNA microarray analysis of rat alveolar epithelial cells following
exposure to organic extract of diesel exhaust particles. Toxicology and applied pharmacology 201: 178-185.
Koike E; Kobayashi T. (2005). Organic extract of diesel exhaust particles stimulates expression of la and costimulatory
molecules associated with antigen presentation in rat peripheral blood monocytes but not in alveolar macrophages.
Toxicology and applied pharmacology 209: 277-285.
Koken PJ; Piver WT; Ye F; Elixhauser A; Olsen LM; Portier CJ. (2003). Temperature, air pollution, and hospitalization for
cardiovascular diseases among elderly people in Denver. Environ Health Perspect 111: 1312-1317.
Kongerud J; Madden MC; Hazucha M; Peden D. (2006). Nasal responses in asthmatic and nonasthmatic subjects following
exposure to diesel exhaust particles. Inhal Toxicol 18: 589-594.
Kooter IM; Boere AJF; Fokkens PHB; Leseman DLAC; Dormans JAMA; Cassee FR. (2006). Response of spontaneously
hypertensive rats to inhalation of fine and ultrafme particles from traffic: experimental controlled study. Particle and fibre
toxicology 3.
Korsnes R; Pavlova O; Godtliebsen F. (2002). Assessment of potential transport of pollutants into the Barents Sea via sea ice—an
observational approach. Marine Pollution Bulletin 44: 861-869.
Koutrakis P; Suh HH; Sarnat JA; Brown KW; Coull BA; Schwartz J. (2005). Characterization of particulate and gas exposures of
sensitive subpopulations living in Baltimore and Boston. ResRepHealth Efflnst: 1-65.
Kreyling WG. (1992). Intracellular particle dissolution in alveolar macrophages. Environmental health perspectives 97: 121-121.
Kreyling WG; Scheuch G. (2000). Clearance of Particles Deposited in the Lungs. LUNG BIOLOGY IN HEALTH AND
DISEASE 143: 323-366.
Kreyling WG; Semmler-Behnke M; Moller W. (2006). Ultrafine particle-lung interactions: does size matter? J Aerosol Med 19:
74-83.
Krieger UK. (2007). Simultaneous Measurements of PM 10 and PM 1 using a single TEOM#. Aerosol Science and Technology
41: 975-980.
Kristjansson JE. (2002). Studies of the aerosol indirect effect from sulfate and black carbon aerosols. Journal of Geophysical
Research 107: 1-19.
Kroll JH; Seinfeld JH. (2008). Chemistry of secondary organic aerosol: Formation and evolution of low-volatility organics in the
atmosphere. Atmospheric Environment 42: 3593-3624.
Kruize H; Hanninen O; Breugelmans O; Lebret E; Jantunen M. (2003). Description and demonstration of the EXPOLIS
simulation model: two examples of modeling population exposure to particulate matter. J ExpoAnalEnvironEpidemiol
13: 87-99.
Kucharski J; Wyszkowska J. (2004). Inter-relationship between number of microorganisms and spring barley yield and degree of
soil contamination with copper. Plant Soil Environ 50: 243-249.
Kuhn T; Biswas S; Sioutas C. (2005a). Diurnal and seasonal characteristics of particle volatility and chemical composition in the
vicinity of a light-duty vehicle freeway. Atmospheric Environment 39: 7154-7166.
Kuhn T; Krudysz M; Zhu Y; Fine PM; Hinds WC; Froines J; Sioutas C. (2005b). Volatility of indoor and outdoor ultrafine
particulate matter near a freeway. J Aerosol Sci 36: 291-302.
Kuhns H; Knipping EM. (2005). Development of a United States-Mexico emissions inventory for the Big Bend Regional Aerosol
and Visibility Observational (BRAVO) Study. Journal of the Air & Waste Management Association 55: 677-692.
Kulkarni MM; Patil RS. (2003). Personal Exposure to Toxic Metals in an Indian Metropolitan Region. Journal of the Institution
of Engineers (India): Environmental Engineering Division.
Kulkarni N; Pierse N; Rushton L; Grigg J. (2006). Carbon in airway macrophages and lung function in children. The New
England journal of medicine 355: 21-30.
Kulmala M; Mordas G; Petaja T; Gronholm T; Aalto PP; Vehkamaki H; Hienola AI; Herrmann E; Sipila M; Riipinen I. (2007).
The condensation particle counter battery (CPCB): A new tool to investigate the activation properties of nanoparticles. J
Aerosol Sci 38: 289-304.
Kumar VS; Mani U; Prasad AK; Lai K; Gowri V; Gupta A. (2004). Effect of fly ash inhalation on biochemical and
histomorphological changes in rat lungs. Indian journal of experimental biology 42: 964-968.
December 2008
R-38
DRAFT-DO NOT QUOTE OR CITE
-------
Kunzli N; Jerrett M; Mack WJ; Beckerman B; LaBree L; Gilliland F; Thomas D; Peters J; Hodis HN. (2005). Ambient air
pollution and atherosclerosis in Los Angeles. Environ Health Perspect 113: 201-206.
Kuo HW; Lai JS; Lee MC; Tai RC; Lee MC. (2002). Respiratory effects of air pollutants among asthmatics in central Taiwan.
Archives of environmental health 57: 194-200.
Kurniawan A; Schmidt-Ott A. (2006). Monitoring the Soot Emissions of Passing Cars. Environmental science & technology 40:
1911-1915.
Kutti T; Ervik A; Lfoisaster T. (2008). Effects of organic effluents from a salmon farm on a fjord system. III. Linking deposition
rates of organic matter and benthic productivity. Aquaculture 282: 47-53.
LaBrecque JJ; Benzo Z; Alfonso JA; Cordoves Manuelita Quintal PR; Gomez CV; Marcano E. (2004). The concentrations of
selected trace elements in clams, Trivela mactroidea along the Venezuelan coast in the state of Miranda. Marine Pollution
Bulletin 49: 659-667.
Lacasana M; Esplugues A; Ballester F. (2005). Exposure to ambient air pollution and prenatal and early childhood health effects.
European journal of epidemiology 20: 183-199.
Laden F; Neas LM; Dockery DW; Schwartz J. (2000). Association of Fine Particulate Matter from Different Sources with Daily
Mortality in Six US Cities. Environmental health perspectives 108: 941-947.
Laden F; Schwartz J; Speizer FE; Dockery DW. (2006). Reduction in fine particulate air pollution and mortality: Extended
follow-up of the Harvard Six Cities study. American journal of respiratory and critical care medicine 173: 667-672.
Lagorio S; Forastiere F; Pistelli R; Iavarone I; Michelozzi P; Fano V; Marconi A; Ziemacki G; Ostro BD. (2006). Air pollution
and lung function among susceptible adult subjects: a panel study. Environ Health 5:11.
LaitinenLA; Heino M; Laitinen A; Kava T; Haahtela T. (1985). Damage of the airway epithelium and bronchial reactivity in
patients with asthma. Am Rev Respir Dis 131: 599-606.
Lake DA; Tolocka MP; Johnston MV; Wexler AS. (2003). Mass Spectrometry of Individual Particles between 50 and 750 nm in
Diameter at the Baltimore Supersite. Environmental science & technology 37: 3268-3274.
Lake DA; Tolocka MP; Johnston MV; Wexler AS. (2004). The character of single particle sulfate in Baltimore. Atmospheric
Environment 38: 5311-5320.
Lakzian A; Murphy P; Turner A; Beynon JL; Giller KE. (2002). Rhizobium leguminosarum bv. viciae populations in soils with
increasing heavy metal contamination: abundance, plasmid profiles, diversity and metal tolerance. Soil Biology &
Biochemistry 34: 519-529.
Lambert MRK. (1997). Environmental effects of heavy spillage from a destroyed pesticide store near Hargeisa (Somaliland)
assessed during the dry season, using reptiles and amphibians as bioindicators. Archives of Environmental Contamination
and Toxicology 32: 80-93.
Landis MS; Keeler GJ. (2002). Atmospheric mercury deposition to Lake Michigan during the Lake Michigan Mass Balance
Study. Environmental Science and Technology 36: 4518-4524.
Landvik NE; Gorria M; Arlt VM; Asare N; Solhaug A; Lagadic-Gossmann D; Holme JA. (2007). Effects of nitrated-polycyclic
aromatic hydrocarbons and diesel exhaust particle extracts on cell signalling related to apoptosis: possible implications
for their mutagenic and carcinogenic effects. Toxicology 231: 159-174.
Lanki T; de Hartog JJ; Heinrich J; Hoek G; Janssen NA; Peters A; Stolzel M; Timonen KL; Vallius M; Vanninen E; Pekkanen J.
(2006b). Can we identify sources of fine particles responsible for exercise-induced ischemia on days with elevated air
pollution? The ULTRA study. Environ Health Perspect 114: 655-660.
Lanki T; Pekkanen J; Aalto P; Elosua R; Berglind N; DTppoliti D; Kulmala M; Nyberg F; Peters A; Picciotto S; Salomaa V;
Sunyer J; Tiittanen P; von Klot S; Forastiere F. (2006a). Associations of traffic related air pollutants with hospitalisation
for first acute myocardial infarction: the HEAPSS study. Occupational and environmental medicine 63: 844-851.
Larrieu S; Jusot JF; Blanchard M; Prouvost H; Declercq C; Fabre P; Pascal L; Tertre AL; Wagner V; Riviere S; Chardon B;
Borrelli D; Cassadou S; Eilstein D; Lefranc A. (2007). Short term effects of air pollution on hospitalizations for
cardiovascular diseases in eight French cities: the PSAS program. The Science of the total environment 387: 105-112.
Larson T; Gould T; Simpson C; Liu LJ; Claiborn C; Lewtas J. (2004). Source apportionment of indoor, outdoor, and personal
PM2.5 in Seattle, Washington, using positive matrix factorization. J Air Waste ManagAssoc 54: 1175-1187.
Larsson BM; Sehlstedt M; Grunewald J; Skold CM; Lundin A; Blomberg A; Sandstrom T; Eklund A; Svartengren M. (2007).
Road tunnel air pollution induces bronchoalveolar inflammation in healthy subjects. Eur Respir J 29: 699-705.
Last JA; Ward R; Temple L; Pinkerton KE; Kenyon NJ. (2004). Ovalbumin-induced airway inflammation and fibrosis in mice
also exposed to ultrafme particles. Inhalation toxicology 16: 93-102.
Laurent O; Pedrono G; Segala C; Filleul L; Havard S; Deguen S; Schillinger C; Riviere E; Bard D. (2008). Air pollution, asthma
attacks, and socioeconomic deprivation: a small-area case-crossover study. Am J Epidemiol 168: 58-65.
Lawson ST; Scherbatskoy TD; Malcolmb EG; Keeler GJ. (2003). Cloud water and throughfall deposition of mercury and trace
elements in a high elevation spruce-fir forest at Mt. Mansfield, Vermont. J Environ Monit 5: 578-583.
Lay JC; Bennett WD; Kim CS; Devlin RB; Bromberg PA. (1998). Retention and intracellular distribution of instilled iron oxide
particles in human alveolar macrophages. American Journal of Respiratory Cell and Molecular Biology 18: 687-695.
Lay JC; Stang MR; Fisher PE; Yankaskas JR; Bennett WD. (2003). Airway Retention of Materials of Different Solubility
following Local Intrabronchial Deposition in Dogs. J Aerosol Med 16: 153-166.
Lay JC; Zeman KL; Ghio AJ; Bennett WD. (2001). Effects of inhaled iron oxide particles on alveolar epithelial permeability in
normal subjects. Inhalation Toxicology 13: 1065-1078.
December 2008
R-39
DRAFT-DO NOT QUOTE OR CITE
-------
Le Tertre A; Medina S; Samoli E; Forsberg B; Michelozzi P; Boumghar A; Vonk JM; Bellini A; Atkinson R; Ayres JG. (2002a).
Short-term effects of particulate air pollution on cardiovascular diseases in eight European cities. 56: 773-779.
Le Tertre A; Medina S; Samoli E; Forsberg B; Michelozzi P; Boumghar A; Vonk JM; Bellini A; Atkinson R; Ayres JG; Sunyer
J; Schwartz J; Katsouyanni K. (2002b). Short-term effects of particulate air pollution on cardiovascular diseases in eight
European cities. Journal of epidemiology and community health 56: 773-779.
Le Tertre A; Medina S; Samoli E; Forsberg B; Michelozzi P; Boumghar A; Vonk JM; Bellini A; Atkinson R; Ayres JG; Sunyer
J; Schwartz J; Katsouyanni K. (2003). Short-term effects of particulate air pollution on cardiovascular diseases in eight
European cities. In Revised Analyses of Time-Series Studies of Air Pollution and Health. Special Report. Boston, MA:
Health Effects Institute.
Le Tertre A; Schwartz J; Touloumi G. (2005). Empirical Bayes and adjusted estimates approach to estimating the relation of
mortality to exposure of PM(10). Risk Anal 25: 711-718.
LeBlanc GA. (1995). Trophic-level differences in the bioconcentration of chemicals: Implications in assessing environmental
biomagnification. Environmental Science and Technology 29: 154-160.
Lee BE; Ha EH; Park HS; Kim YJ; Hong YC; Kim H; Lee JT. (2003a). Exposure to air pollution during different gestational
phases contributes to risks of low birth weight. Hum Reprod 18: 638-643.
Lee CC; Cheng YW; Kang JJ. (2005a). Motorcycle exhaust particles induce IL-8 production through NF-kappaB activation in
human airway epithelial cells. Journal of toxicology and environmental health part A 68: 1537-1555.
Lee HM. (2007). Fabrication of Reference Filter for Measurements of EC (Elemental Carbon) and OC (Organic Carbon) in
Aerosol Particles. Aerosol Science and Technology 41: 284-294.
Lee HM; Kim CS; Shimada M; Okuyama K. (2005c). Effects of Mobility Changes and Distribution of Bipolar Ions on Aerosol
Nanoparticle Diffusion Charging. Journal of Chemical Engineering of Japan 38: 486-496.
Lee HM; Soo Kim C; Shimada M; Okuyama K. (2005b). Bipolar diffusion charging for aerosol nanoparticle measurement using
a soft X-ray charger. J Aerosol Sci 36: 813-829.
Lee JH; Hopke PK; Holsen TM; Lee DW; A. Jaques P; Sioutas C; Ambs JL. (2005d). Performance evaluation of continuous
PM2. 5 mass concentration monitors. J Aerosol Sci 36: 95-109.
Lee JT; Kim H; Cho YS; Hong YC; Ha EH; Park H. (2003b). Air pollution and hospital admissions for ischemic heart diseases
among individuals 64+ years of age residing in Seoul, Korea. Archives of environmental health 58: 617-623.
Lee JT; Kim H; Song H; Hong YC; Cho YS; Shin SY; Hyun YJ; Kim YS. (2002). Air pollution and asthma among children in
Seoul, Korea. Epidemiology 13: 481-484.
Lee JT; Son JY; Cho YS. (2007). The adverse effects of fine particle air pollution on respiratory function in the elderly. The
Science of the total environment 385: 28-36.
Lee KP; Henry Iii NW; Trochimowicz HJ; Reinhardt CF. (1986). Pulmonary response to impaired lung clearance in rats
following excessive TiO sub (2) dust deposition. Environmental Research 41: 144-167.
Lee KP; Trochimowicz HJ; Reinhardt CF. (1985a). Pulmonary response of rats exposed to titanium dioxide (Ti02) by inhalation
for two years. Toxicology and applied pharmacology 79: 179-192.
Lee KP; Trochimowicz HJ; Reinhardt CF. (1985b). Transmigration of titanium dioxide (Ti02) particles in rats after inhalation
exposure. Exp Mol Pathol 42: 331-343.
Lee SJ; Delgado-Saborit JM; Demokritou P; Koutrakis P. (2006a). Development and evaluation of personal respirable particulate
sampler (PRPS). Atmospheric Environment 40: 212-224.
Lee SL; Wong WH; Lau YL. (2006b). Association between air pollution and asthma admission among children in Hong Kong.
Clin Exp Allergy 36: 1138-1146.
Lee SS; Woo CH; Chang JD; Kim JH. (2003c). Roles of Rac and cytosolic phospholipase A2 in the intracellular signalling in
response to titanium particles. Cellular Signalling 15: 339-345.
Lee T; Yu XY; Ayres B; Kreidenweis SM; Malm WC; Collett JL. (2008). Observations of fine and coarse particle nitrate at
several rural locations in the United States. Atmospheric Environment 42: 2720-2732.
Lee WY; Iannucci-Berger WA; Eitzer BD; White JC; Mattina MI. (2003d). Plant uptake and translocation of air-borne chlordane
and comparison with the soil-to-plant route. Chemosphere 53: 111-121.
Leem JH; Kaplan BM; Shim YK; Pohl HR; Gotway CA; Bullard SM; Rogers JF; Smith MM; Tylenda CA. (2006). Exposures to
air pollutants during pregnancy and preterm delivery. Environ Health Perspect 114: 905-910.
Lei Y-C; Chan C-C; Wang P-Y; Lee C-T; Cheng T-J. (2004a). Effects of Asian dust event particles on inflammation markers in
peripheral blood and bronchoalveolar lavage in pulmonary hypertensive rats. Environmental Research 95: 71-76.
Lei YC; Chen MC; Chan CC; Wang PY; Lee CT; Cheng TJ. (2004b). Effects of concentrated ambient particles on airway
responsiveness and pulmonary inflammation in pulmonary hypertensive rats. Inhal Toxicol 16: 785-792.
Lei YD; Wania F. (2004). Is rain or snow a more efficient scavenger of organic chemicals? Atmospheric Environment 38: 3557-
3571.
Leith D; Sommeriatt D; Boundy MG. (2007). Passive Sampler for PM 10-2. 5 Aerosol. Journal of the Air & Waste Management
Association 57.
Lemos M; Mohallen SV; Macchione M; Dolhnikoff M; AssunA§A£o JoV; Godleski JJ; Saldiva PHN. (2006). Chronic Exposure
to Urban Air Pollution Induces Structural Alterations in Murine Pulmonary and Coronary Arteries. Inhalation Toxicology
18: 247-253.
Letz AG; Quinn JM. (2005). Relationship of basic military trainee emergency department visits for asthma and San Antonio air
quality. Allergy Asthma Proc 26: 463-467.
December 2008
R-40
DRAFT-DO NOT QUOTE OR CITE
-------
Levin SA. (1998). Ecosystems and the biosphere as complex adaptive systems. Ecosystems 1: 431-436.
LeVine AM; Gwozdz J; Stark J; Bruno M; Whitsett J; Korfhagen T. (1999). Surfactant protein-A enhances respiratory syncytial
virus clearance in vivo. J Clin Invest 103: 1015-1021.
LeVine AM; Whitsett JA. (2001). Pulmonary collectins and innate host defense of the lung. Microbes Infect 3: 161-166.
Levy D; Sheppard L; Checkoway H; Kaufman J; Lumley T; Koenig J; Siscovick D. (2001). A case-crossover analysis of
particulate matter air pollution and out-of-hospital primary cardiac arrest. Epidemiology 12: 193-199.
Levy JI; Wilson AM; Evans JS; Spengler JD. (2003). Estimation of primary and secondary particulate matter intake fractions for
power plants in Georgia. Environ Sci Technol 37: 5528-5536.
Lewis TC; Robins TG; Dvonch JT; Keeler GJ; Yip FY; Mentz GB; Lin X; Parker EA; Israel BA; Gonzalez L; Hill Y. (2005). Air
pollution-associated changes in lung function among asthmatic children in Detroit. Environ Health Perspect 113: 1068-
1075.
Lewne M; Nise G; Lind ML; Gustavsson P. (2006). Exposure to particles and nitrogen dioxide among taxi, bus and lorry drivers.
Int ArchOccup Environ Health 79: 220-226.
Li N; Nel AE. (2006). Role of the Nrf2-mediated signaling pathway as a negative regulator of inflammation: Implications for the
impact of particulate pollutants on asthma. Antioxidants & Redox Signaling 8: 88-98.
Li N; Sioutas C; Cho A; Schmitz D; Misra C; Sempf J; Wang M; Oberley T; Froines J; Nel A. (2003). Ultrafme particulate
pollutants induce oxidative stress and mitochondrial damage. Environmental health perspectives 111: 455-460.
Li Y-F; MacDonald RW; Jantunen LMM; Harner T; Bidleman TF; Stachen WMJ. (2002). The transport of P-
hexachlorocyclohexane to the western Arctic Ocean: a contrast to aHCH. The Science of the total environment 291: 229-
246.
Li Y-J; Kawada T; Matsumoto A; Azuma A; Kudoh S; Takizawa H; Sugawara I. (2007). Airway inflammatory responses to
oxidative stress induced by low-dose diesel exhaust particle exposure differ between mouse strains. Experimental lung
research 33: 227-244.
Li Z; Hyseni X; Carter ID; Soukup JM; Dailey LA; Huang Y-CT. (2006). Pollutant particles enhanced H202 production from
NAD(P)H oxidase and mitochondria in human pulmonary artery endothelial cells. American journal of physiology Cell
physiology 291: C357-365.
Liao D; Duan Y; Whitsel EA; Zheng Z-j; Heiss G; Chinchilli VM; Lin H-M. (2004). Association of Higher Levels of Ambient
Criteria Pollutants with Impaired Cardiac Autonomic Control: A Population-based Study. Am J Epidemiol 159: 768-777.
Liao D; Heiss G; Chinchilli VM; Duan Y; Folsom AR; Lin H-M; Salomaa V. (2005). Association of criteria pollutants with
plasma hemostatic/inflammatory markers: a population-based study. Journal of exposure analysis and environmental
epidemiology 15: 319-328.
Librando V; Tringali G. (2005). Atmospheric fate of OH initiated oxidation of terpenes. Reaction mechanism of alpha-pinene
degradation and secondary organic aerosol formation. J Environ Manage 75: 275-282.
Lichtenfels AJFC; Gomes JB; Pieri PC; Miraglia SGEK; Hallak J; Saldiva PHN. (2007). Increased levels of air pollution and a
decrease in the human and mouse male-to-female ration in Sao Paulo, Brazil. Fertility and sterility 87: 230-232.
Liebhart J; Malolepszy J; Wojtyniak B; Pisiewicz K; Plusa T; Gladysz U. (2007). Prevalence and risk factors for asthma in
Poland: results from the PMSEAD study. J InvestigAllergolClinlmmunol 17: 367-374.
Lim HJ; Allen G; Maring H; Solomon P; Turpin BJ; Edgerton E; Hering SV. (2003). Semicontinuous aerosol carbon
measurements: Comparison of Atlanta supersite measurements. Journal of Geophysical Research D: Atmospheres 108:
SOS-SOS.
Lim L; Wurl O; Karuppiah S; Obbard JP. (2007a). Atmospheric wet deposition of PAHs to the sea-surface micro layer. Marine
Pollution Bulletin 54: 1212-1219.
Lim MCH; Ayoko GA; Morawska L; Ristovski ZD; Jayaratne ER. (2007b). The effects of fuel characteristics and engine
operating conditions on the elemental composition of emissions from heavy duty diesel buses. Fuel 86: 1831-1839.
Lin CA; Pereira LA; Nishioka DC; Conceicao GM; Braga AL; Saldiva PH. (2004a). Air pollution and neonatal deaths in Sao
Paulo, Brazil. Braz J Med Biol Res 37: 765-770.
Lin CC; Chang CT; Li TC; Kao A. (2002a). Objective Evidence of Impairment of Alveolar Integrity inpatients withNon-
Insulin-Dependent Diabetes Mellitus Using Radionuclide Inhalation Lung Scan. Lung 180: 181-186.
Lin CM; Li CY; Mao IF. (2004b). Increased risks of term low-birth-weight infants in a petrochemical industrial city with high air
pollution levels. Archives of environmental health 59: 663-668.
Lin H; Tao S; Zuo Q; Coveney RM. (2007). Uptake of polycyclic aromatic hydrocarbons by maize plants. Environmental
Pollution 148: 614-619.
Lin M; Chen Y; Burnett RT; Villeneuve PJ; Krewski D. (2002b). The influence of ambient coarse particulate matter on asthma
hospitalization in children: case-crossover and time-series analyses. Environ Health Perspect 110: 575-581.
Lin M; Stieb DM; Chen Y. (2005). Coarse particulate matter and hospitalization for respiratory infections in children younger
than 15 years in Toronto: a case-crossover analysis. Pediatrics 116: e235-240.
Lindberg SE; Lovett GM. (1985). Field measurements of particle dry deposition rates to foliage and inert surfaces in a forest
canopy. Environmental Science and Technology 19: 238-244.
Lindbom J; Gustafsson M; Blomqvist G; Dahl A; Gudmundsson A; Swietlicki E; Ljungman AG. (2007). Wear particles
generated from studded tires and pavement induces inflammatory reactions in mouse macrophage cells. Chemical
Research in toxicology 20: 937-946.
December 2008
R-41
DRAFT-DO NOT QUOTE OR CITE
-------
Lingua G; Franchin C; Todeschini V; Castiglione S; S.Biondi; Burlando B; Parravicini V; P.Torrigiani; G.Berta. (2008).
Arbuscular mycorrhizal fungi differentially affect the response to high zinc concentrations of two registered poplar
clones. Environmental Pollution 153: 137-147.
Linn WS; Gong H, Jr. (1999). The 21st century environment and air quality influences on asthma. Curr Opin Pulm Med 5: 21-26.
Linn WS; Szlachcic Y; Gong H, Jr.; Kinney PL; Berhane KT. (2000). Air pollution and daily hospital admissions in metropolitan
Los Angeles. Environ Health Perspect 108: 427-434.
Lipfert F; Wyzga R; Baty J; Miller J. (2006a). Traffic density as a surrogate measure of environmental exposures in studies of air
pollution health effects: Long-term mortality in a cohort of US veterans. Atmospheric Environment 40: 154-169.
Lipfert FW; Baty JD; Miller JP; Wyzga RE. (2006b). PM2.5 constituents and related air quality variables as predictors of
survival in a cohort of U.S. military veterans. Inhal Toxicol 18: 645-657.
Lipfert FW; Morris SC. (2002). Temporal and spatial relations between age specific mortality and ambient air quality in the
United States: regression results for counties, 1960-97. Occupational and environmental medicine 59: 156-174.
Lipfert FW; Zhang J; Wyzga RE. (2000). Infant mortality and air pollution: a comprehensive analysis of U.S. data 1990. J Air
Waste Manag Assoc 50: 1350-1366.
Lippmann M; Gordon T; Chen Lung C. (2005a). Effects of subchronic exposures to concentrated ambient particles in mice. IX.
Integral assessment and human health implications of subchronic exposures of mice to CAPs. Inhalation toxicology 17:
255-261.
Lippmann M; Hwang J-S; Maciejczyk P; Chen L-C. (2005b). PM source apportionment for short-term cardiac function changes
in ApoE-/-mice. Environmental health perspectives 113: 1575-1579.
Lippmann M; Ito K; Hwang JS; Maciejczyk P; Chen LC. (2006). Cardiovascular effects of nickel in ambient air. Environ Health
Perspect 114: 1662-1669.
Lippmann M; Ito K; Nadas A; Burnett RT. (2000). Association of particulate matter components with daily mortality and
morbidity in urban populations. Res Rep Health Eff Inst: 5-72, discussion 73-82.
Lipsett MJ; Tsai FC; Roger L; Woo M; Ostro BD. (2006). Coarse particles and heart rate variability among older adults with
coronary artery disease in the Coachella Valley, California. Environ Health Perspect 114: 1215-1220.
Lisabeth LD; Escobar JD; Dvonch JT; Sanchez BN; Majersik JJ; Brown DL; Smith MA; Morgenstern LB. (2008). Ambient air
pollution and risk for ischemic stroke and transient ischemic attack. Annals of neurology 64: 53-59.
Little P; Wiffen RD. (1977). Emission and deposition of petrol engine exhaust Pb—I. Deposition of exhaust Pb to plant and soil
surfaces. Atmospheric Environment 11: 437-447.
Liu HH; Wu YC; Chen HL. (2007a). Production of ozone and reactive oxygen species after welding. ArchEnviron Contam
Toxicol 53: 513-518.
Liu J; Ballaney M; Al-Alem U; Quan C; Jin X; Perera F; Chen LC; Miller RL. (2008). Combined Inhaled Diesel Exhaust
Particles and Allergen Exposure Alter Methylation of T Helper Genes andlgE Production In Vivo. Toxicol Sci 102: 76-
81.
Liu L; Ruddy TD; Dalipaj M; Szyszkowicz M; You H; Poon R; Wheeler A; Dales R. (2007b). Influence of personal exposure to
particulate air pollution on cardiovascular physiology and biomarkers of inflammation and oxidative stress in subjects
with diabetes. J Occup Environ Med 49: 258-265.
Liu LJ; Box M; Kalman D; Kaufman J; Koenig J; Larson T; Lumley T; Sheppard L; Wallace L. (2003). Exposure assessment of
particulate matter for susceptible populations in Seattle. EnvironHealthPerspect 111: 909-918.
Liu LJ; Cuijuric I; Keidel D; Heldstab J; Kunzli N; Bayer-Oglesby L; Ackermann-Liebrich U; Schindler C. (2007c).
Characterization of source-specific air pollution exposure for a large population-based Swiss cohort (SAPALDIA).
Environ Health Perspect 115: 1638-1645.
Liu PL; Chen YL; Chen YH; Lin SJ; Kou YR. (2005a). Wood smoke extract induces oxidative stress-mediated caspase-
independent apoptosis in human lung endothelial cells: role of AIF and EndoG. Am J Physiol Lung Cell Mol Physiol 289:
L739-749.
Liu S; Krewski D; Shi Y; Chen Y; Burnett RT. (2007d). Association between maternal exposure to ambient air pollutants during
pregnancy and fetal growth restriction. Journal of Exposure Science and Environmental Epidemiology 17: 426-432.
Liu W; Zhang L; Weisel CP; Turpin B; Morandi M; Stock T; Colome S; Zhang J; Korn LR. (2007e). Predicting personal
exposure to airborne carbonyls using residential measurements and time/activity data. Atmospheric Environment Indoor
Air 2005 - 10th International Conference on Indoor Air Quality and Climate (Part II) 41: 5280-5288.
Liu X; Hegg DA; Stoelinga MT. (2001). Numerical simulation of new particle formation over the northwest Atlantic using the
MM5 mesoscale model coupled with sulfur chemistry (Paper 2000JD900765). JOURNAL OF GEOPHYSICAL
RESEARCH-ALL SERIES- 106: 9697-9716.
Liu X; Meng Z. (2005). Effects of airborne fine particulate matter on antioxidant capacity and lipid peroxidation in multiple
organs of rats. Inhalation toxicology 17: 467-473.
Liu Y. (2005). Atmospheric response and feedback to radiative forcing from biomass burning in tropical South America.
Agricultural and Forest Meteorology 133: 40-53.
Liu Y-Q; Keane M; Ensell M; Miller W; Kashon M; Ong T-m; Mauderly J; Lawson D; Gautam M; Zielinska B; Whitney K;
Eberhardt J; Wallace W. (2005b). In vitro genotoxicity of exhaust emissions of diesel and gasoline engine vehicles
operated on a unified driving cycle. J Environ Monit 7: 60-66.
Llorca J; Salas A; Prieto-Salceda D; Chinchon-Bengoechea V; Delgado-Rodriguez MCEH. (2005). Nitrogen dioxide increases
cardiorespiratory admissions in Torrelavega (Spain). J Environ Health 68: 30-35.
December 2008
R-42
DRAFT-DO NOT QUOTE OR CITE
-------
Lofroth G. (1981). Salmonella/microsome mutagenicity assay of exhaust from diesel and gas-powered vehicles. Env Int 5: 255-
261.
Lohman U; Feichter J. (2005). Global indirect aerosol effects: a review. Atmospheric Chemistry and Physics 5: 715-737.
Long CM; Suh HH; Kobzik L; Catalano PJ; Ning YY; Koutrakis P. (2001). A pilot investigation of the relative toxicity of indoor
and outdoor fine particles: in vitro effects of endotoxin and other particulate properties. EnvironHealth Perspect 109:
1019-1026.
Long JF; Waldman WJ; Kristovich R; Williams M; Knight D; Dutta PK. (2005). Comparison of ultrastructural cytotoxic effects
of carbon and carbon/iron particulates on human monocyte-derived macrophages. Environ Health Perspect 113: 170-174.
Long RW; McClenny WA. (2006). Laboratory and field evaluation of instrumentation for the semicontinuous determination of
particulate nitrate (and other water-soluble particulate components). J Air Waste Manag Assoc 56: 294-305.
Loomis D; Castillejos M; Gold DR; McDonnell W; Borja-Aburto VH. (1999). Air pollution and infant mortality in Mexico City.
Epidemiology 10: 118-123.
Lopez Alonso M; Benedito JL; Miranda M; Castillo C; Hernandez J; Shore RF. (2002). Cattle as biomonitors of environmental
semi-metal and trace metal concentrations in Galicia (NW Spain). Archives of Environmental Contamination and
Toxicology 43: 103-108.
Lopez Alonso M; Benedito JL; Miranda M; Castillo C; Hernandez J; Shore RF. (2003b). Mercury concentrations in cattle from
NW Spain. Science of the Total Environment 302.
Lopez Alonso M; Benedito JL; Miranda M; Fernandez JA; Castillo C; Hernandez J; Shore RF. (2003a). Large-scale spatial
variation in mercury concentrations in cattle in NW Spain. Environmental Pollution 125: 173-181.
Lorino AM; Meignan M; Bouissou P; Atlan G. (1989). Effects of sustained exercise on pulmonary clearance of aerosolized
99mTc-DTPA. Journal of applied physiology 67: 2055-2059.
Lorusso S; Marabelli M; Troili M. (1997). Air pollution and the deterioration of historic monuments. Journal of Environmental
Pathology, Toxicology, and Oncology 16: 171-173.
Lovett GM. (1994). Atmospheric deposition of nutrients and pollutants in North America: an ecological perspective. Ecological
Applications 4(4): 629-650.
Lowenthal DH; Kumar N. (2003). PM sub (2. 5) Mass and Light Extinction Reconstruction in IMPROVE. Journal of the Air &
Waste Management Association 53: 1109-1120.
Lu R; Turco RP; Jacobson MZ. (1997). An integrated air pollution modeling system for urban and regional scales: 1. Structure
and performance. Journal of Geophysical Research D Atmospheres 102: 6063-6079.
Lu R; Turco RP; Stolzenbach K; Friedlander SK; Xiong C; Schiff K; Tiefenthaler L; Wang G. (2003). Dry deposition of airborne
trace metals on the Los Angeles Basin and adjacent coastal waters. Journal of Geophysical Research D: Atmospheres
108.
Lubac B; Loisel H; Guiselin N; Astoreca R; Artigas LF; Meriaux X. (2008). Hyperspectral and multispectral ocean color
inversions to detect Phaeocystis globosa blooms in coastal waters. Journal of Geophysical Research-Oceans 113.
Lucaciu A; Timofte L; Culicov O; Frontasyeva MV; Oprea C; Cucu-Man S; Mocanu R; Steinnes E. (2004). Atmospheric
deposition of trace elements in Romania studied by the moss biomonitoring technique. J Atmos Chem 49: 533-548.
Luginaah IN; Fung KY; Gorey KM; Webster G; Wills C. (2005). Association of ambient air pollution with respiratory
hospitalization in a government-designated "area of concern": the case of Windsor, Ontario. Environ Health Perspect 113:
290-296.
Lunackova L; Masarovicova E; Kraova K; Stresko V. (2003). Response of fast growing woody plants from family Salicaceae to
cadmium treatment. Bulletin of Environmental Contamination and Toxicology 70: 576-585.
Lund AK; Knuckles TL; Obot AC; Shohet R; McDonald JD; Gigliotti A; Seagrave JC; Campen MJ. (2007). Gasoline exhaust
emissions induce vascular remodeling pathways involved in atherosclerosis. Toxicological sciences 95: 485-494.
Lunden MM; Kirchstetter TW; Thatcher TL; Hering SV; Brown NJ. (2008). Factors affecting the indoor concentrations of
carbonaceous aerosols of outdoor origin. Atmospheric Environment 42: 5660-5671.
Lunden MM; Littlejohn D; Hering SV; Brown NJ; Revzan KL; Fischer ML; Thatcher TL. (2003a). The transformation of
outdoor ammonium nitrate aerosols in the indoor environment. Atmospheric Environment 37: 5633-5644.
Lunden MM; Thatcher TL; Hering SV; Brown NJ. (2003b). Use of Time-and Chemically Resolved Particulate Data To
Characterize the Infiltration of Outdoor PM~ 2~.~ 5 into a Residence in the San Joaquin Valley. ENVIRONMENTAL
SCIENCE AND TECHNOLOGY-WASHINGTON DC- 37: 4724-4732.
Lundgren DL; Hahn FF; Crain CR; Sanchez A. (1978). Effect of Influenza Virus Infection on the Pulmonary Retention of
Inhaled 144Ce and Subsequent Survival of Mice. Health Physics 34: 557.
Luo C; Mahowald NM; Del Corral J. (2003). Sensitivity study of meteorological parameters on mineral aerosol mobilization,
transport, and distribution. Journal of Geophysical Research 108 (D15), 4447, doi:10.1029/2003JD003483.
Luttmann-Gibson H; Suh H; Coull BA; Dockery DW; Sarnat SE; Schwartz J; Stone PH; Gold DR. (2006). Short-term effects of
air pollution on heart rate variability in senior adults in Steubenville, Ohio. J Occup Environ Med 48: 780-788.
Lynam MM; Keeler GJ. (2005). Artifacts associated with the measurement of particulate mercury in an urban environment: the
influence of elevated ozone concentrations. Atmospheric Environment 39: 3081-3088.
Maciejczyk P; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice. VIII. Source-
related daily variations in in vitro responses to CAPs. Inhalation toxicology 17: 243-253.
December 2008
R-43
DRAFT-DO NOT QUOTE OR CITE
-------
Maciejczyk P; Zhong M; Li Q; Xiong J; Nadziejko C; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient
particles (CAPs) in mice. II. The design of a CAPs exposure system for biometric telemetry monitoring. Inhalation
toxicology 17: 189-197.
Mackay D. (1991). Multimedia Environmental Models: The Fugacity Approach. Chelsea: Lewis Publishers.
MackmanN. (2005). Tissue-Specific Hemostasis in Mice. 25: 2273-2281.
Mackman N; Tilley RE; Key NS. (2007). Role of the extrinsic pathway of blood coagulation in hemostasis and thrombosis.
Arterioscler Thromb Vase Biol 27: 1687-1693.
Macleod M; McKone TE; Mackay D. (2005). Mass balance for mercury in the San Francisco Bay area. Environmental Science
and Technology 39: 6721-6729.
Mader BT; Schauer JJ; Seinfeld JH; Flagan RC; Yu JZ; Yang H; Lim HJ; Turpin BJ; Deminter JT; Heidemann G. (2003).
Sampling methods used for the collection of particle-phase organic and elemental carbon during ACE-Asia. Atmospheric
Environment 37: 1435-1449.
Madrid L; Diaz-Barrientos E; Madrid F. (2002). Distribution of heavy metal contents of urban soils in parks of Seville.
Chemosphere 49: 1301-1309.
Magas OK; Gunter JT; Regens JL. (2007). Ambient air pollution and daily pediatric hospitalizations for asthma. Environ Sci
Pollut Res Int 14: 19-23.
Maheswaran R; Haining RP; Brindley P; Law J; Pearson T; Fryers PR; Wise S; Campbell MX (2005a). Outdoor air pollution,
mortality, and hospital admissions from coronary heart disease in Sheffield, UK: a small-area level ecological study.
European heart journal 26: 2543-2549.
Maheswaran R; Haining RP; Brindley P; Law J; Pearson T; Fryers PR; Wise S; Campbell MJ. (2005b). Outdoor air pollution and
stroke in Sheffield, United Kingdom: a small-area level geographical study. Stroke; a journal of cerebral circulation 36:
239-243.
Mahlman ID. (1997). Uncertainties in projections of human-caused climate warming. Science 278: 1416-1417.
Mahowald N; Zender C; Luo C; Savoie D; Torres O; del Corral J. (2002). Under- understanding the 30-year Barbados desert dust
record. Journal of Geophysical Research 107.
Maisonet M; Bush TJ; Correa A; Jaakkola JJ. (2001). Relation between ambient air pollution and low birth weight in the
Northeastern United States. Environ Health Perspect 109 Suppl 3: 351-356.
Maisonet M; Correa A; Misra D; Jaakkola JJ. (2004). A review of the literature on the effects of ambient air pollution on fetal
growth. Environ Res 95: 106-115.
Makar PA; Gravel S; Chirkov V; Strawbridge KB; Froude F; Arnold J; Brook J. (2006). Heat flux, urban properties, and regional
weather. Atmospheric Environment 40: 2750-2766.
Maletto A; McKendry IG; Strawbridge KB. (2003). Profiles of particulate matter size distributions using a balloon-borne
lightweight aerosol spectrometer in the planetary boundary layer. Atmospheric Environment 37: 661-670.
Malkonen E; Derome J; Fritze H; Helmisaari H-S; Kukkola M; Kyto M; Saarsalmi A; Salemaa M. (1999). Compensatory
fertilization of Scots pine stands polluted by heavy metals. Nutrient Cycling in Agroecosy stems 55: 239-268.
Malm WC; Collett, Jr.; McMeeking G; Lee T; Carrillo J; Schichtel B; Day DE; Carrico C; Kreidenweis SM. (2005).
Intercomparison and closure calculations using measurements of aerosol species and optical properties during the
Yosemite aerosol characterization study. Journal of Geophysical Research D: Atmospheres 110: 1-21.
Malm WC; Hand JL. (2007). An examination of the physical and optical properties of aerosols collected in the IMPROVE
program. Atmospheric Environment 41: 3407-3427.
Malm WC; National Park S; United S; Air; Water Quality D. (1983). Introduction to Visibility: National Park Service, Air and
Water Quality Division, Air Research Branch.
Malm WC; Pitchford ML; McDade C; Ashbaugh LL. (2007). Coarse particle speciation at selected locations in the rural
continental United States. Atmospheric Environment 41: 2225-2239.
Malm WC; Schichtel BA; Ames RB; Gebhart KA. (2002). A Ten-Year Spatial and Temporal Trend of Sulfate across the United
States. J Geophys Res.
Malm WC; Schichtel BA; Pitchford ML; Ashbaugh LL; Eldred RA. (2004). Spatial and monthly trends in speciated fine particle
concentration in the United States. J Geophys Res 109: 3306.
Mangum JB; Bermudez E; Sar M; Everitt JL (2004). Osteopontin expression in particle-induced lung disease. Experimental lung
research 30: 585-598.
Mannes T; Jalaludin B; Morgan G; Lincoln D; Sheppeard V; Corbett S. (2005). Impact of ambient air pollution on birth weight in
Sydney, Australia. Occupational and environmental medicine 62: 524-530.
Mar TF; Ito K; Koenig JQ; Larson TV; Eatough DJ; Henry RC; Kim E; Laden F; Lall R; Neas L; Stolzel M; Paatero P; Hopke
PK; Thurston GD. (2006). PM source apportionment and health effects. 3. Investigation of inter-method variations in
associations between estimated source contributions of PM2.5 and daily mortality in Phoenix, AZ. Journal of Exposure
Science and Environmental Epidemiology 16: 311-320.
Mar TF; Jansen K; Shepherd K; Lumley T; Larson TV; Koenig JQ. (2005a). Exhaled nitric oxide in children with asthma and
short-term PM2.5 exposure in Seattle. Environ Health Perspect 113: 1791-1794.
Mar TF; Koenig JQ; Jansen K; Sullivan J; Kaufman J; Trenga CA; Siahpush SH; Liu LJ; Neas L. (2005b). Fine particulate air
pollution and cardiorespiratory effects in the elderly. Epidemiology 16: 681-687.
Mar TF; Larson TV; Stier RA; Claiborn C; Koenig JQ. (2004). An analysis of the association between respiratory symptoms in
subjects with asthma and daily air pollution in Spokane, Washington. Inhal Toxicol 16: 809-815.
December 2008
R-44
DRAFT-DO NOT QUOTE OR CITE
-------
Mar TF; Norris GA; Koenig JQ; Larson TV. (2000). Associations between Air Pollution and Mortality in Phoenix, 1995-1997.
Environmental health perspectives 108: 347-353.
Mar TF; Norris GA; Larson TV; Wilson WE; Koenig JQ. (2003). Air pollution and cardiovascular mortality in Phoenix, 1995—
1997. Health Effects Institute, Cambridge, MA.
Marinoni N; Birelli MP; Rostagno C; Pavese A. (2003). The effects of atmospheric multipollutants on modern concrete.
Atmospheric Environment 37: 4701-4712.
Markiewicz Patkowska J; Hursthouse A; Przybyla-Kij H. (2005). The interaction of heavy metals with urban soils: sorption
behaviour of Cd, Cu, Cr, Pb and Zn with a typical mixed brownfield deposit. Environment International 31: 513-521.
Markovic S; Mitric M; Starcevic G; Uskokovic D. (2008). Ultrasonic de-agglomeration of barium titanate powder. Ultrason
Sonochem 15: 16-20.
Martin MH; Coughtrey PJ. (1981). Impact of metals on ecosystem function and productivity. In Lepp NW (Ed.), Effect of heavy
metal pollution on plants: volume 2, metals in the environment (pp. 119-158). Barking, United Kingdom: Applied
Science Publishers.
Martin RV; Chance K; Jacob DJ; Kurosu TP; Spurr RJD; Bucsela E; Gleason JF; Palmer PI; Bey I; Fiore AM. (2002). An
improved retrieval of tropospheric nitrogen dioxide from GOME. J Geophys Res 107: 9-1.
Martins LC; Latorre Mdo R; Cardoso MR; Goncalves FL; Saldiva PH; Braga AL. (2002). Air pollution and emergency room
visits due to pneumonia and influenza in Sao Paulo, Brazil. Rev Saude Publica 36: 88-94.
Martins MA; Santos C; Almeida MM; Costa ME. (2008). Hydroxyapatite micro- and nanoparticles: nucleation and growth
mechanisms in the presence of citrate species. J Colloid Interface Sci 318: 210-216.
Martins MC; Fatigati FL; Vespoli TC; Martins LC; Pereira LA; Martins MA; Saldiva PH; Braga AL. (2004). Influence of
socioeconomic conditions on air pollution adverse health effects in elderly people: an analysis of six regions in Sao
Paulo, Brazil. Journal of epidemiology and community health 58: 41-46.
Martonen TB; Schroeter JD. (2003). Risk assessment dosimetry model for inhaled particulate matter: II. Laboratory surrogates
(rat). Toxicology letters 138: 133-142.
Martonen TB; Yang Y; Xue ZQ. (1994). Effects of Carinal Ridge Shapes on Lung Airstreams. Aerosol Science and Technology
21: 119-136.
Masclet PP; Hoyau V; Jaffrezo JL; Cachier H. (2000). Polycyclic aromatic hydrocarbons deposition on the ice sheet of
Greenland. Parti. Superficial snow. Atmospheric Environment 34: 3195-3207.
Masjedi MR; Jamaati HR; Dokouhaki P; Ahmadzadeh Z; Taheri SA; Bigdeli M; Izadi S; Rostamian A; Aagin K; Ghavam SM.
(2003). The effects of air pollution on acute respiratory conditions. Respirology 8: 213-230.
Mason GR; Uszler JM; Effros RM; Reid E. (1983). Rapidly reversible alterations of pulmonary epithelial permeability induced
by smoking. Chest 83: 6-11.
Massicotte R; Robidoux PY; Sauve S; Flipo D; Fournier M; Trottier B. (2003). Immune response of earthworms (Lumbricus
terrestris, Eisenia andrei and Aporrectodea tuberculata) following in situ soil exposure to atmospheric deposition from a
cement factory. J Environ Monit 5: 774-779.
Massie ST; Heymsfield A; Schmitt C; Muller D; Seifert P. (2007). Aerosol indirect effects as a function of cloud top pressure.
Journal of Geophysical Research 112: 6202.
Mathis U; Mohr M; Forss AM. (2005). Comprehensive particle characterization of modern gasoline and diesel passenger cars at
low ambient temperatures. Atmospheric Environment 39: 107-117.
Mathur R. (2008). Estimating the impact of the 2004 Alaskan forest fires on episodic particulate matter pollution over the eastern
United States through assimilation of satellite-derived aerosol optical depths in a regional air quality model. J Geophys
Res 113.
Matsui H; Randell SH; Peretti SW; William Davis C; Boucher RC. (1998). Coordinated Clearance of Periciliary Liquid and
Mucus from Airway Surfaces. Journal of clinical investigation 102: 1125-1131.
Matsumoto A; Hiramatsu K; Li Y; Azuma A; Kudoh S; Takizawa H; Sugawara I. (2006). Repeated exposure to low-dose diesel
exhaust after allergen challenge exaggerates asthmatic responses in mice. Clinical immunology 121: 227-235.
Matsumoto K; Hayano T; Uematsu M. (2003). Positive artifact in the measurement of particulate carbonaceous substances using
an ambient carbon particulate monitor. Atmospheric Environment 37: 4713-4717.
Matthias I, et al. (2004). Multiyear aerosol observations with dual wavelength Raman lidar in the framework of EARLINET.
Journal of Geophysical Research 109, D13203, doi:10.1029/2004JD004600.
Matthias-Maser S; Bogs B; Jaenicke R. (2000). The size distribution of primary biological aerosol particles in cloud water on the
mountain Kleiner Feldberg/Taunus (FRG). Atmospheric Research 54: 1-13.
Matti Maricq M. (2007). Chemical characterization of particulate emissions from diesel engines: A review. J Aerosol Sci 38:
1079-1118.
Mauad T; Rivero DH; de Oliveira RC; Lichtenfels AJ; Guimaraes ET; de Andre PA; Kasahara DI; Bueno HM; Saldiva PH.
(2008). Chronic exposure to ambient levels of urban particles affects mouse lung development. American journal of
respiratory and critical care medicine 178: 721-728.
Mauderly JL. (1997). Relevance of particle-induced rat lung tumors for assessing lung carcinogenic hazard and human lung
cancer risk. Environmental health perspectives 105: 1337-1346.
Mayo JM; Legge AH; Yeung EC; Krupa SV; Bogner JC. (1992). The effects of sulphur gas and elemental sulphur dust
deposition on Pinus contorta x Pinus banksiana: Cell walls and water relations. Environmental Pollution 76: 43-50.
December 2008
R-45
DRAFT-DO NOT QUOTE OR CITE
-------
McBride SJ; Williams RW; Creason J. (2007). Bayesian hierarchical modeling of personal exposure to particulate matter.
Atmospheric Environment 41: 6143-6155.
McConnell JR; Edwards R; Kok GL; Flanner MG; Zender CS; Saltzman ES; Banta JR; Pasteris DR; Carter MM; Kahl DWJ.
(2007). 20th-century industrial black carbon emissions altered arctic climate forcing. Science 317: 1381-1384.
McConnell R; Berhane K; Gilliland F; London SJ; Vora H; Avol E; Gauderman WJ; Margolis HG; Lurmann F; Thomas DC;
Peters JM. (1999). Air pollution and bronchitic symptoms in Southern California children with asthma. Environ Elealth
Perspect 107: 757-760.
McConnell R; Berhane K; Gilliland F; Molitor J; Thomas D; Lurmann F; Avol E; Gauderman WJ; Peters JM. (2003).
Prospective study of air pollution and bronchitic symptoms in children with asthma. American journal of respiratory and
critical care medicine 168: 790-797.
McCormack MC; Breysse PN; Hansel NN; Matsui EC; Tonorezos ES; Curtin-Brosnan J; Williams DL; Buckley TJ; Eggleston
PA; Diette GB. (2007). Common household activities are associated with elevated particulate matter concentrations in
bedrooms of inner-city Baltimore pre-school children. EnvironRes.
McCreanor J; Cullinan P; Nieuwenhuijsen MJ; Stewart-Evans J; Malliarou E; Jarup L; Harrington R; Svartengren M; Han IK;
Ohman-Strickland P; Chung KF; Zhang J. (2007). Respiratory effects of exposure to diesel traffic in persons with asthma.
The New England journal of medicine 357: 2348-2358.
McCurdy T; Glen G; Smith L; Lakkadi Y. (2000). The national exposure research laboratory's consolidated human activity
database. Journal of exposure analysis and environmental epidemiology 10: 566-578.
McDonald JD; Eide I; Seagrave J; Zielinska B; Whitney K; Lawson DR; Mauderly JL. (2004). Relationship between
composition and toxicity of motor vehicle emission samples. Environmental health perspectives 112: 1527-1538.
McDonald JD; Reed MD; Campen MJ; Barrett EG; Seagrave J; Mauderly JL. (2007). Health effects of inhaled gasoline engine
emissions. Inhal Toxicol 19 Suppl 1: 107-116.
McDonnell WF; Nishino-Ishikawa N; Petersen FF; Chen LH; Abbey DE. (2000). Relationships of mortality with the fine and
coarse fractions of long-term ambient PM10 concentrations in nonsmokers. Journal of exposure analysis and
environmental epidemiology 10: 427-436.
McEntee JC; Ogneva-Himmelberger Y. (2008). Diesel particulate matter, lung cancer, and asthma incidences along major traffic
corridors in MA, USA: A GIS analysis. Health Place 14: 817-828.
McFiggans A; Midgley PM. (2001). The effect of aerosol composition and properties on warm cloud droplet activation.
Atmospheric Chemistry and Physics 6: 2593-2649.
McGowan TF; Lipinski GE; Santoleri JJ. (1993). New rules affect the handling of waste fuels. Chemical Engineering (New
York): 122-128.
McKendry IG; Strawbridge KB; O'Neill NT; Macdonald AM; Liu PSK; Leaitch WR; Anlauf KG; Jaegle L; Fairlie TD;
Westphal DL. (2007). Trans-Pacific transport of Saharan dust to western North America: A case study. J Geophys Res
112.
McQueen DS; Donaldson K; Bond SM; McNeilly JD; Newman S; Barton NJ; Duffin R. (2007). Bilateral vagotomy or atropine
pre-treatment reduces experimental diesel-soot induced lung inflammation. Toxicology and applied pharmacology 219:
62-71.
Mebust MR; Eder BK; Binkowski FS; Roselle SJ. (2003). Models-3 community multiscale air quality (CMAQ) model aerosol
component 2. Model evaluation. J Geophys Res 108: 4184.
Medeiros A; Gouveia N. (2005). [Relationship between low birthweight and air pollution in the city of Sao Paulo, Brazil]. Rev
Saude Publica 39: 965-972.
Medeiros N, Jr.; Rivero DH; Kasahara DI; Saiki M; Godleski JJ; Koutrakis P; Capelozzi VL; Saldiva PH; Antonangelo L.
(2004). Acute pulmonary and hematological effects of two types of particle surrogates are influenced by their elemental
composition. Environmental research 95: 62-70.
Medina-Ramon M; Zanobetti A; Schwartz J. (2006). The effect of ozone and PM10 on hospital admissions for pneumonia and
chronic obstructive pulmonary disease: a national multicity study. Am J Epidemiol 163: 579-588.
Meers E; Vangronsveld J; Tack FMG; Vandecasteele B; Ruttens A. (2007). Potential of five willow species (Salix spp.) for
phytoextraction of heavy metals. Environmental and Experimental Botany 60: 57-68.
Mehta RH; Anad Kumar TC. (1997). Declining semen quality in Bangaloreans: A preliminary report. Current science(Bangalore)
72: 621-622.
Meignan M; Rosso J; Leveau J; Katz A; Cinotti L; Madelaine G; Galle P. (1986). Exercise increases the lung clearance of
inhaled technetium-99m DTPA. J Nucl Med 27: 274-280.
Melnikov S; Carroll J; Gorshkov A; Vlasov S; Dahle S. (2003). Snow and ice concentrations of selected persistent organic
pollutants in the Ob-Yenisey River watershed. The Science of the total environment 306: 27-37.
Menache MG; Miller FJ; Raabe OG. (1995). Particle inhalability curves for humans and small laboratory animals. Ann Occup
Hyg 39: 317-328.
Meng QY; Turpin BJ; Korn L; Weisel CP; Morandi M; Colome S; Zhang J; Stock T; Spektor D; Winer A. (2004). Influence of
ambient (outdoor) sources on residential indoor and personal PM2. 5 concentrations: Analyses of RIOPA data. Journal of
Exposure Science and Environmental Epidemiology 15: 17.
December 2008
R-46
DRAFT-DO NOT QUOTE OR CITE
-------
Meng QY; Turpin BJ; Korn L; Weisel CP; Morandi M; Colome S; Zhang JJ; Stock T; Spektor D; Winer A; Zhang L; Lee JH;
Giovanetti R; Cui W; Kwon J; Alimokhtari S; Shendell D; Jones J; Farrar C; Maberti S. (2005a). Influence of ambient
(outdoor) sources on residential indoor and personal PM2.5 concentrations: analyses of RIOPA data. J
ExpoAnalEnvironEpidemiol 15: 17-28.
Meng QY; Turpin BJ; Lee JH; Polidori A; Weisel CP; Morandi M; Colome S; Zhang J; Stock T; Winer A. (2007a). How does
infiltration behavior modify the composition of ambient PM2.5 in indoor spaces? An analysis of RIOPA data. EnvironSci
Technol 41: 7315-7321.
Meng QY; Turpin BJ; Polidori A; Lee JH; Weisel C; Morandi M; Colome S; Stock T; Winer A; Zhang J. (2005b). PM2.5 of
ambient origin: estimates and exposure errors relevant to PM epidemiology. EnvironSci Technol 39: 5105-5112.
Meng YY; Wilhelm M; Rull RP; English P; Ritz B. (2007b). Traffic and outdoor air pollution levels near residences and poorly
controlled asthma in adults. Ann Allergy Asthma Immunol 98: 455-463.
Meng Z; Dabdub D; Seinfeld JH. (1997). Chemical Coupling Between Atmospheric Ozone and Particulate Matter. Science 277:
116.
Menon S. (2004). Current uncertainties in assessing aerosol effects on climate. Annual Review of Environment and Resources
29: 1-30.
Mensink C; De Ridder K; Deutsch F; Lefebre F; Van de Vel K. (2008). Examples of scale interactions in local, urban, and
regional air quality modelling. Atmospheric Research 89: 351-357.
Metzger KB; Klein M; Flanders WD; Peel JL; Mulholland JA; Langberg JJ; Tolbert PE. (2007). Ambient air pollution and
cardiac arrhythmias in patients with implantable defibrillators. Epidemiology 18: 585-592.
Metzger KB; Tolbert PE; Klein M; Peel JL; Flanders WD; Todd K; Mulholland JA; Ryan PB; Frumkin H. (2004). Ambient air
pollution and cardiovascular emergency department visits. Epidemiology 15: 46-56.
Mfula AM; Kukadia V; Griffiths RF; Hall DJ. (2005). Wind tunnel modelling of urban building exposure to outdoor pollution.
Atmospheric Environment 39: 2737-2745.
Michaud JP; Grove JS; Krupitsky DCEH. (2004). Emergency department visits and "vog"-related air quality in Hilo, Hawai'i.
Environ Res95: 11-19.
Middha P; Wexler A. (2006). Design of a Slot Nanoparticle Virtual Impactor. Aerosol Science and Technology 40: 737-743.
Middlebrook AM; Murphy DM; Lee SH; Thomson DS; Prather KA; Wenzel RJ; Liu DY; Phares DJ; Rhoads KP; Wexler AS.
(2003). A comparison of particle mass spectrometers during the 1999 Atlanta Supersite Project. J Geophys Res 108:
8424.
MiddletonN; Yiallouros P; Kleanthous S; Kolokotroni O; Schwartz J; Dockery DW; Demokritou P; Koutrakis P. (2008). A 10-
year time-series analysis of respiratory and cardiovascular morbidity in Nicosia, Cyprus: the effect of short-term changes
in air pollution and dust storms. Environ Health 7: 39.
Middleton WEK. (1952). Vision Through the Atmosphere.
Mie G. (1908). Beitrage zur Optik trixber Medien, speziell kolloidaler Metallosungen. Annalen der Physik 330: 377-445.
Migliaretti G; Cadum E; Migliore E; Cavallo F. (2005). Traffic air pollution and hospital admission for asthma: a case-control
approach in a Turin (Italy) population. Int Arch Occup Environ Health 78: 164-169.
Migliaretti G; Cavallo F. (2004). Urban air pollution and asthma in children. Pediatr Pulmonol 38: 198-203.
Miller AL; Habjan MC; Park K. (2007a). Real-time estimation of elemental carbon emitted from a diesel engine. Environ Sci
Technol 41: 5783-5788.
Miller FJ. (2000). Dosimetry of particles in laboratory animals and humans in relationship to issues surrounding lung overload
and human health risk assessment: a critical review. Inhal Toxicol 12: 19-57.
Miller FJ; Gardner DE; Graham JA; Lee Jr RE; Wilson WE; Bachmann JD. (1979). Size considerations for establishing a
standard for inhalable particles. J Air Pollut Control Assoc 29: 610-615.
Miller KA; Siscovick DS; Sheppard L; Shepherd K; Sullivan JH; Anderson GL; Kaufman JD. (2007b). Long-term exposure to
air pollution and incidence of cardiovascular events in women. The New England journal of medicine 356: 447-458.
Miller RL; Tegen I; Perlwitz J. (2004). Surface radiative forcing by soil dust aerosols and the hydrologic cycle. Journal of
Geophysical Research 109.
Miller SL; Anderson MJ; Daly EP; Milford JB. (2002). Source apportionment of exposures to volatile organic compounds. I.
Evaluation of receptor models using simulated exposure data. Atmospheric Environment 36: 3629-3641.
Mills NL; Amin N; Robinson SD; Anand A; Davies J; Patel D; de la Fuente JM; Cassee FR; Boon NA; Macnee W; Millar AM;
Donaldson K; Newby DE. (2006). Do inhaled carbon nanoparticles translocate directly into the circulation in humans?
American j ournal of respiratory and critical care medicine 173: 426-431.
Mills NL; Robinson SD; Fokkens PH; Leseman DL; Miller MR; Anderson D; Freney EJ; Heal MR; Donovan RJ; Blomberg A;
Sandstrom T; MacNee W; BoonNA; Donaldson K; Newby DE; Cassee FR. (2008). Exposure to concentrated ambient
particles does not affect vascular function in patients with coronary heart disease. Environ Health Perspect 116: 709-715.
Mills NL; Tornqvist H; Gonzalez MC; Vink E; Robinson SD; Soderberg S; Boon NA; Donaldson K; Sandstrom T; Blomberg A;
Newby DE. (2007). Ischemic and Thrombotic Effects of Dilute Diesel-Exhaust Inhalation in Men with Coronary Heart
Disease. The New England journal of medicine 357: 1075-1082.
Mills NL; Tornqvist H; Robinson SD; Gonzalez M; Darnley K; MacNee W; Boon NA; Donaldson K; Blomberg A; Sandstrom T;
Newby DE. (2005). Diesel exhaust inhalation causes vascular dysfunction and impaired endogenous fibrinolysis.
Circulation 112: 3930-3936.
December 2008
R-47
DRAFT-DO NOT QUOTE OR CITE
-------
Mircea M, et al. (2005). Importance of the organic aerosol fraction for modeling aerosol hygroscopic growth and activation: a
case study in the Amazon Basin. Atmospheric Chemistry and Physics 5: 3111-3126.
Miyazaki Y; Kondo Y; Takegawa N; Komazaki Y; Fukuda M; Kawamura K; Mochida M; Okuzawa K; Weber RJ. (2006). Time-
resolved measurements of water-soluble organic carbon in Tokyo. J Geophys Res 111.
Mogensen CE. (1984). Microalbuminuria predicts clinical proteinuria and early mortality in maturity-onset diabetes. The New
England journal of medicine 310: 356-360.
Mohallem SV; Ja de Araujo Lobo D; Regina Pesquero C; Vicente Assunjao J; Afonso de Andre P; Hilario Nascimento Saldiva
P; Dolhnikoff M. (2005). Decreased fertility in mice exposed to environmental air pollution in the city of Sao Paulo.
Environmental Research 98: 196-202.
Molinelli AR; Madden MC; McGee JK; Stonehuerner JG; Ghio AJ. (2002). Effect of metal removal on the toxicity of airborne
particulate matter from the Utah Valley. Inhal Toxicol 2002: 1069-1086.
Moller W; Brown DM; Kreyling WG; Stone V. (2005). Ultrafine particles cause cytoskeletal dysfunctions in macrophages: role
of intracellular calcium. PartFibreToxicol 2: 7.
Moller W; Felten K; Sommerer K; Scheuch G; Meyer G; Meyer P; Elaussinger K; Kreyling WG. (2008). Deposition, retention,
and translocation of ultrafine particles from the central airways and lung periphery. American journal of respiratory and
critical care medicine 177: 426-432.
Moller W; Elaussinger K; Winkler-Heil R; Stahlhofen W; Meyer T; Elofmann W; Eleyder J. (2004). Mucociliary and long-term
particle clearance in the airways of healthy nonsmoker subjects. Journal of applied physiology 97: 2200-2206.
MolletNR; Cademartiri F; de Feyter PJ. (2005). Non-invasive multislice CT coronary imaging. Eleart 91: 401-407.
Molnar P; Bellander T; Sallsten G; Boman J. (2007). Indoor and outdoor concentrations of PM2.5 trace elements at homes,
preschools and schools in Stockholm, Sweden. J EnvironMonit 9: 348-357.
Molnar P; Johannesson S; Boman J; Barregard L; Sallsten G. (2006). Personal exposures and indoor, residential outdoor, and
urban background levels of fine particle trace elements in the general population. J EnvironMonit 8: 543-551.
Monn C. (2001). Exposure assessment of air pollutants: a review on spatial heterogeneity and indoor/outdoor/personal exposure
to suspended particulate matter, nitrogen dioxide and ozone. Atmospheric Environment 35: 1-32.
Monn C; Naef R; Roller T. (2003). Reactions of macrophages exposed to particles <10 microm. Environ Res 91: 35-44.
Montauban van Swijndregt AD; De Lange EE; De Groot E; Ackerstaff RG. (1999). An in vivo evaluation of the reproducibility
of intima-media thickness measurements of the carotid artery segments using B-mode ultrasound. Ultrasound Med Biol
25: 323-330.
Moolgavkar SH. (2000). Air pollution and hospital admissions for diseases of the circulatory system in three U.S. metropolitan
areas. J Air Waste Manag Assoc 50: 1199-1206.
Moore KF; Ning Z; Ntziachristos L; Schauer JJ; Sioutas C. (2007). Daily variation in the properties of urban ultrafine aerosol—
Part I: Physical characterization and volatility. Atmospheric Environment 41: 8633-8646.
Moore KJ; Kunjathoor VV; Koehn SL; Manning JJ; Tseng AA; Silver JM; McKee M; Freeman MW. (2005). Loss of receptor-
mediated lipid uptake via scavenger receptor A or CD36 pathways does not ameliorate atherosclerosis in hyperlipidemic
mice. J Clin Invest 115: 2192-2201.
Morawska L; Keogh DU; Thomas SB; Mengersen K. (2007). Modality in ambient particle size distributions and its potential as a
basis for developing air quality regulation. Atmospheric Environment.
Morawska L; Thomas S; Jamriska M; Johnson G. (1999). The modality of particle size distributions of environmental aerosols.
Atmospheric Environment 33: 4401-4411.
Morgenstern V; Zutavern A; Cyrys J; Brockow I; Koletzko S; Kramer U; Behrendt H; Herbarth O; von Berg A; Bauer CP;
Wichmann HE; Heinrich J. (2008). Atopic diseases, allergic sensitization, and exposure to traffic-related air pollution in
children. American journal of respiratory and critical care medicine 177: 1331-1337.
Morishita M; Keeler G; Wagner J; Marsik F; Timm E; Dvonch J; Harkema J. (2004). Pulmonary retention of particulate matter is
associated with airway inflammation in allergic rats exposed to air pollution in urban Detroit. Inhalation toxicology 16:
663-674.
Moropoulou A; BisbikouK; Torfs K; Van GriekenR; ZezzaF; Macri F. (1998). Origin and growth of weathering crusts on
ancient marbles in industrial atmosphere. Atmospheric Environment 32: 967-982.
Morris RD; Naumova EN. (1998). Carbon monoxide and hospital admissions for congestive heart failure: evidence of an
increased effect at low temperatures. Environ Health Perspect 106: 649-653.
Morrow PE. (1994). Mechanisms and Significance of "Particle Overload" In: Toxic and Carcinogenic Effects of Solid Particles
in the Respiratory Tract. Edited by U. Mohr: ILSI Press Washington DC.
Mortimer K; Neugebauer R; Lurmann F; Alcorn S; Balmes J; Tager I. (2008). Air pollution and pulmonary function in asthmatic
children: effects of prenatal and lifetime exposures. Epidemiology 19: 550-557; discussion 561-552.
Mortimer KM; Neas LM; Dockery DW; Redline S; Tager IB. (2002). The effect of air pollution on inner-city children with
asthma. EurRespir J 19: 699-705.
Moses CA; Smith BJ. (1994). Limestone weathering in the supratidal zone: an example from Mallorca. In Robinson DA, G.
WRB (Eds.), Rock weathering and landform evolution (pp. 432^151). Chichester: John Wiley & Sons.
Moshammer H; Hutter HP; Hauck H; Neuberger M. (2006). Low levels of air pollution induce changes of lung function in a
panel of schoolchildren. Eur Respir J 27: 1138-1143.
Moshammer H; Neuberger M. (2003). The active surface of suspended particles as a predictor of lung function and pulmonary
symptoms in Austrian school children. Atmospheric Environment 37: 1737-1744.
December 2008
R-48
DRAFT-DO NOT QUOTE OR CITE
-------
Mousa K; Onadeko BO; Mustafa HT; Mohamed M; Nabilla A; Omar A; Al-Bunni A; Elgazzar A. (2000). Technetium 99mTc-
DTPA clearance in the evaluation of pulmonary involvement in patients with diabetes mellitus. Respir Med 94: 1053-
1056.
Muggenburg BA; Barr EB; Cheng YS; Seagrave JC; Tilley LP; Mauderley JL. (2000). Effect of inhaled residual oil fly ash on
the electrocardiogram of dogs. Inhal Toxicol 12 Suppl 4: 189-208.
Muhle H; Creutzenberg O; Bellmann B; Heinrich U; Mermelstein R. (1990). Dust overloading of lungs: Investigations of various
materials, species differences, and irreversibility of effects. J Aerosol Med 3:111-128.
Muir DCG; Omelchenko A; Grift NP; Savoie DA; Lockhart WL; Wilkinson P; Brunskill GJ. (1996). Spatial trends and historical
deposition of polychlorinated biphenyls in Canadian midlatitude and arctic lake sediments. Environmental Science and
Technology 30: 3609-3617.
Muller K; Spindler G; Maenhaut W; Hitzenberger R; Wieprecht W; Baltensperger U; ten Brink H. (2004). INTERCOMP2000, a
campaign to assess the comparability of methods in use in Europe for measuring aerosol composition. Atmospheric
Environment 38: 6459-6466.
Mumford JL; He XZ; Chapman RS; Harris DB; Li XM; Xian YL; Jiang WZ; Xu CW; Chuang JC; et. (1987). Lung cancer and
indoor air pollution in Xuan Wei, China. Science 235: 217-220.
Murahashi T. (2003). Determination of mutagenic 3-nitrobenzanthrone in diesel exhaust particulate matter by three-dimensional
high-performance liquid chromatography. The Analyst 128: 42-45.
Murata A; Kida K; Hasunuma H; Kanegae H; Ishimaru Y; Motegi T; Yamada K; Yoshioka H; Yamamoto K; Kudoh S. (2007).
Environmental influence on the measurement of exhaled nitric oxide concentration in school children: special reference
to methodology. J Nippon Med Sch 74: 30-36.
Murayama T, et al. (2001). Ground-based network observation of Asian dust events of April 1998 in East Asia. Journal of
Geophysical Research 106: 18345-18360.
Murray M; Holmes SA. (2004). Assessment of mercury emissions inventories for the Great Lakes states. Environmental
Research 95: 282-297.
Mutlu GM; Green D; Bellmeyer A; Baker CM; Burgess Z; Rajamannan N; Christman JW; Foiles N; Kamp DW; Ghio AJ;
Chandel NS; Dean DA; Sznajder JI; Budinger GRS. (2007). Ambient particulate matter accelerates coagulation via an IL-
6-dependent pathway. Journal of clinical investigation 117: 2952-2961.
Mutlu GM; Snyder C; Bellmeyer A; Wang H; Hawkins K; Soberanes S; Welch LC; Ghio AJ; Chandel NS; Kamp D; Sznajder
Jacob I; Budinger GRS. (2006). Airborne particulate matter inhibits alveolar fluid reabsorption in mice via oxidant
generation. American journal of respiratory cell and molecular biology 34: 670-676.
Nadel JA; Salem H; Tamplin B; Tokiwa Y. (1965a). Mechanism of bronchoconstriction during inhalation of sulfur dioxide.
Journal of applied physiology 20: 164-167.
Nadel JA; Tamplin B; Tokiwa Y. (1965b). Mechanism of Bronchoconstriction. During Inhalation of Sulfur Dioxide; Reflex
Involving Vagus Nerves. Archives of environmental health 10: 175-178.
Nadziejko C; Fang K; Nadziejko E; Narciso SP; Zhong M; Chen LC. (2002). Immediate effects of particulate air pollutants on
heart rate and respiratory rate in hypertensive rats. Cardiovasc Toxicol 2: 245-252.
Nadziejko C; Fang K; Narciso S; Zhong M; Su Wei C; Gordon T; Nadas A; Chen Lung C. (2004). Effect of particulate and
gaseous pollutants on spontaneous arrhythmias in aged rats. Inhalation Toxicology 16: 373-380.
Naess O, Piro FN, Nafstad P, Smith GD, Leyland AH. 2007. Air pollution, social deprivation, and mortality: a
multilevel cohort study. Epidemiology 18(6): 686-694.
Naess O; Nafstad P; Aamodt G; Claussen B; Rosland P. (2007a). Relation between concentration of air pollution and cause-
specific mortality: four-year exposures to nitrogen dioxide and particulate matter pollutants in 470 neighborhoods in
Oslo, Norway. Am J Epidemiol 165: 435-443.
Naess O; Piro FN; Nafstad P; Smith GD; Leyland AH. (2007b). Air pollution, social deprivation, and mortality: a multilevel
cohort study. Epidemiology 18: 686-694.
Nakamura K; Nakawo M; Ageta Y; Goto-Azuma K; Kamiyam K. (2000). Post-depositional loss of nitrate in surface snow layers
of the Antarctic ice sheet. Bulletin of Glaciology Research 17: 11-16.
NAPA. (1969). Air Quality Criteria for Particulate Matter. National Air Pollution Control Administration, Washington, D.C.
AP-49.
NAPAP. (1991). The experience and legacy of NAPAP. Report of the Oversight Review Board. : National Acid Precipitation
Program., Washington, DC.
Nascimento LF; Pereira LA; Braga AL; Modolo MC; Carvalho JA, Jr. (2006). [Effects of air pollution on children's health in a
city in Southeastern Brazil], Rev Saude Publica 40: 77-82.
Nash DG; Tolocka MP; Baer T. (2006). The uptake of 03 by myristic acid-oleic acid mixed particles: evidence for solid surface
layers. PhysChemChemPhys 8: 4468-4475.
Nash TH, III. (1975). Influence of effluents from a zinc factory on lichens. Ecological Monographs 45: 183-198.
National Kidney Foundation (2008). The National Kidney Foundation Kidney Disease Outcomes Quality Initiative (NKF
KDOQI). Chronic Kidney Disease: Evaluation, Classification, and Stratification. Journal. Retrieved from
http://www.kidney.org/professionals/kdoqi/index.cfm.
Nazaroff WW; Cass GR. (1991). Protecting museum collections from soiling due to the deposition of airborne particles.
Atmospheric Environment 25A: 841-852.
December 2008
R-49
DRAFT-DO NOT QUOTE OR CITE
-------
NCHS. (2007, October 15, 2008). Classification of Diseases, Functioning, and Disability...Monitoring the Nation's Health, from
http://www.cdc.gOv/nchs/icd9.htm#RTF
Neas LM; Schwartz J; Dockery D. (1999). A case-crossover analysis of air pollution and mortality in Philadelphia.
Environmental health perspectives 107: 629-631.
Nemmar A; Al-Maskari S; Ali Badreldin H; Al-Amri Issa S. (2007). Cardiovascular and lung inflammatory effects induced by
systemically administered diesel exhaust particles in rats. American journal of physiology - Lung cellular and molecular
physiology 292: L664-L670.
Nemmar A; Hoet PH; Dinsdale D; Vermylen J; Hoylaerts MF; Nemery B. (2003c). Diesel exhaust particles in lung acutely
enhance experimental peripheral thrombosis. Circulation 107: 1202-1208.
Nemmar A; Hoet PHM; Vanquickenborne B; Dinsdale D; Thomeer M; Hoylaerts MF; Vanbilloen H; Mortelmans L; Nemery B.
(2002). Passage of Inhaled Particles Into the Blood Circulation in Humans. Circulation 105: 411-414.
Nemmar A; Hoylaerts MF; Hoet PH; Vermylen J; Nemery B. (2003b). Size effect of intratracheally instilled particles on
pulmonary inflammation and vascular thrombosis. Toxicol Appl Pharmacol 186: 38-45.
Nemmar A; Inuwa IM. (2008). Diesel exhaust particles in blood trigger systemic and pulmonary morphological alterations.
Toxicology letters 176: 20-30.
Nemmar A; Nemery B; Hoet PHM; Vermylen J; Hoylaerts MF. (2003a). Pulmonary inflammation and thrombogenicity caused
by diesel particles in hamsters: role of histamine. Am J Respir Crit Care Med 168: 1366-1372.
Nerriere E; Guegan H; Bordigoni B; Hautemaniere A; Momas I; Ladner J; Target A; Lameloise P; Delmas V; Personnaz MB;
Koutrakis P; Zmirou-Navier D. (2007). Spatial heterogeneity of personal exposure to airborne metals in French urban
areas. Sci Total Environ 373: 49-56.
Neuberger M; Schimek MG; Horak F; Moshammer H; Kundi M; Frischer T; Gomiscek B; Puxbaum H; Hauck H; Auphep-Team.
(2004). Acute effects of particulate matter on respiratory diseases, symptoms and functions: epidemiological results of the
Austrian Project on Health Effects of Particulate Matter (AUPHEP). Atmospheric Environment 38: 3971-3981.
Newman AB; Sutton-Tyrrell K; Vogt MT; Kuller LH. (1993). Morbidity and mortality in hypertensive adults with a low
ankle/arm blood pressure index. JAMA 270: 487-489.
Ng SP; Kendall M; Dimitroulopoulou C; Grossinho A; Chen LC. (2005). PM2.5 exposure assessment of the population in Lower
Manhattan area of New York City after the World Trade Center disaster. Atmospheric Environment 39: 1979-1992.
Nguyen HL; Leermakers M; Elskens M; De Ridder F; Doan TH; Baeyens W. (2005). Correlations, partitioning and
bioaccumulation of heavy metals between different compartments of Lake Balaton. Science of the Total Environment
341: 211-226.
Nicholson L; Benn DI. (2006). Calculating ice melt beneath a debris layer using meteorological data. Journal of Glaciology 52:
463-470.
Nieminen T; Derome J; Helmisaari H-S. (1999). Interactions between precipitation and Scots pine canopies along a heavy-metal
pollution gradient. Environmental Pollution 106: 129-137.
Nightingale JA; Maggs R; Cullinan P; Donnelly LE; Rogers DF; Kinnersley R; Chung KF; Barnes PJ; Ashmore M; Newman-
Taylor A. (2000). Airway inflammation after controlled exposure to diesel exhaust particulates. Am J Respir Crit Care
Med 162: 161-166.
Niinimaa V; Cole P; Mintz S; Shephard RJ. (1981). Oronasal distribution of respiratory airflow. Respir Physiol 43: 69-75.
Nikolov MC; Coull BA; Catalano PJ; Diaz E; Godleski JJ. (2008). Statistical methods to evaluate health effects associated with
major sources of air pollution: a case-study of breathing patterns during exposure to concentrated Boston air particles.
Journal of the Royal Statistical Society: Series C (Applied Statistics) 57: 357-378.
Nikula KJ; Vallyathan V; Green FHY; Hahn FF. (2001). Influence of Exposure Concentration or Dose on the Distribution of
Particulate Material in Rat and Human Lungs. Environmental health perspectives 109: 311-318.
Ning Z; GellerMD; Moore KF; Sheesley R; Schauer JJ; Sioutas C. (2007). Daily Variation in Chemical Characteristics of Urban
Ultrafme Aerosols and Inference of Their Sources. Environ Science Technol 41: 6000-6006.
Nordenhall C; Pourazar J; Ledin MC; Levin JO; Sandstrom T; Adelroth E. (2001). Diesel exhaust enhances airway
responsiveness in asthmatic subjects. European Respiratory Journal 17: 909-915.
Nordling E; Berglind N; Melen E; Emenius G; Hallberg J; Nyberg F; Pershagen G; Svartengren M; Wickman M; Bellander T.
(2008). Traffic-related air pollution and childhood respiratory symptoms, function and allergies. Epidemiology 19: 401-
408.
Norton SA. (2007). Atmospheric metal pollutants-archives, methods, and history. Water, Air, and Soil Pollution: Focus 7: 93-98.
Noullett M; Jackson PL; Brauer M. (2006). Winter measurements of children's personal exposure and ambient fine particle mass,
sulphate and light absorbing components in a northern community. Atmospheric Environment 40: 1971-1990.
Novakov T; Ramanathan V; Hansen JE; Kirchestetter TW; Sato M. (2003). Large historical changes of fossil-fuel black carbon
aerosols. Geophysical Research Letters 30: 1324.
NRC. (2004). Research priorities for airborne particulate matter. IV. Continuing research progress. Committee on Research
Priorities for Airborne Particulate Matter; Board on Environmental Studies and Toxicology; Commission on Life
Sciences; Commission on Geosciences, Environment, and Resources; National Research Council (NRC); Washington,
DC: National Academies Press.
NRPB. (2004). Particle deposition in the vicinity of power lines and possible effects on health: report of an independent advisory
group on non-ionising radiation and its ad hoc group on corona ions. National Radialogical Protection Board,
December 2008
R-50
DRAFT-DO NOT QUOTE OR CITE
-------
Ntziachristos L; Samaras Z. (2006). Combination of aerosol instrument data into reduced variables to study the consistency of
vehicle exhaust particle measurements. Atmospheric Environment 13th International Symposium on Transport and Air
Pollution (TAP 2004) 40: 6032-6042.
Nurkiewicz TR; Porter DW; Barger M; Castranova V; Boegehold MA. (2004). Particulate matter exposure impairs systemic
microvascular endothelium-dependent dilation. Environmental health perspectives 112: 1299-1306.
Nurkiewicz TR; Porter DW; Barger M; Millecchia L; Rao KMK; Marvar PJ; Hubbs AF; Castranova V; Boegehold MA. (2006).
Systemic microvascular dysfunction and inflammation after pulmonary particulate matter exposure. Environmental health
perspectives 114: 412-419.
Nurkiewicz TR; Porter DW; Hubbs AF; Cumpston JL; Chen BT; Frazer DG; Castranova V. (2008). Nanoparticle inhalation
augments particle-dependent systemic microvascular dysfunction. Part Fibre Toxicol 5: 1.
Nygaard UC; Ormstad H; Aase A; Lovik M. (2005). The IgE adjuvant effect of particles: characterisation of the primary cellular
response in the draining lymph node. Toxicology 206: 181-193.
Nygaard UC; Samuelsen M; Aase A; Lovik M. (2004). The capacity of particles to increase allergic sensitization is predicted by
particle number and surface area, not by particle mass. Toxicological sciences 82: 515-524.
O'Connor PJ; Smith SE; Smith FA. (2002). Arbuscular mycorrhizas influence plant diversity and community structure in a
semiarid herbland. New Phytologist 154: 209-218.
O'Neill MS; Diez-Roux AV; Auchincloss AF1; Franklin TG; Jacobs Jnr DR; Astor BC; Dvonch JT; Kaufman J. (2007). Airborne
particulate matter exposure and urinary albumin excretion: The Multi-Ethnic Study of Atherosclerosis. Occupational and
environmental medicine.
Oberdorster G. (1988). Lung clearance of inhaled insoluble and soluble particles. J Aerosol Med 1: 289-330.
Oberdorster G. (1995). Lung Particle Overload: Implications for Occupational Exposures to Particles. Regulatory toxicology and
pharmacology 21: 123-135.
Oberdorster G. (1996). Significance of Particle Parameters in the Evaluation of Exposure-Dose-Response Relationships of
Inhaled Particles. Inhalation Toxicology 8: 73-89.
Oberdorster G. (2002). Toxicokinetics and Effects of Fibrous andNonfibrous Particles. Inhalation Toxicology 14: 29-56.
Oberdorster G; Ferin J; GeleinR; Soderholm SC; Finkelstein J. (1992). Role of the alveolar macrophage in lung injury: studies
with ultrafine particles. Environmental health perspectives 97: 193-193.
Oberdorster G; Ferin J; Lehnert BE. (1994a). Correlation between particle size, in vivo particle persistence, and lung injury.
Environmental health perspectives 102: 173.
Oberdorster G; Ferin J; Soderholm S; Gelein R; Cox C; Baggs R; Morrow PE. (1994b). Increased pulmonary toxicity of inhaled
ultrafine particles: due to lung overload alone. Ann Occup Hyg 38: 295-302.
Oberdorster G; Finkelstein JN; Johnston C; Gelein R; Cox C; Baggs R; Elder AC. (2000). Acute pulmonary effects of ultrafine
particles in rats and mice. Res Rep Health Eff Inst: 5-74; disc 75-86.
Oberdorster G; Sharp Z; Atudorei V; Elder A; Gelein R; Kreyling W; Cox C. (2004). Translocation of inhaled ultrafine particles
to the brain. Inhalation toxicology 16: 437-445.
Odum EP. (1985). Trends expected in stressed ecosystems. Bioscience 35: 419-422.
Odum EP. (1993). The ecosystem. In: Ecology and our endangered life-support systems (2nd ed.). Sunderland, MA: Sinauer
Associates, Inc.
Oehme M; Biseth A; Schlabach M; Wiig 0. (1995). Concentrations of polychlorinated dibenzo-p-dioxins, dibenzofurans and
non-ortho substituted biphenyls in polar bear milk from Svalbard (Norway). Environmental Pollution 90: 401-407.
Oehme M; Haugen J-E; Schlabach M. (1996). Seasonal changes and relation between levels of organochlorines in Arctic ambient
air: First results from an all-round-year monitoring program at Ny-Alesund, Norway. Environmental Science and
Technology 30: 2294-2304.
Oei HH; Vliegenthart R; Hak AE; Iglesias del Sol A; Hofman A; Oudkerk M; Witteman JC. (2002). The association between
coronary calcification assessed by electron beam computed tomography and measures of extracoronary atherosclerosis:
the Rotterdam Coronary Calcification Study. J Am Coll Cardiol 39: 1745-1751.
Offenberg JH; Lewis CW; Lewandowski M; Jaoui M; Kleindienst TE; Edney EO. (2007). Contributions of toluene and alpha-
pinene to SOA formed in an irradiated toluene/alpha-pinene/NO(x)/ air mixture: comparison of results using 14C content
and SOA organic tracer methods. Environ Sci Technol 41: 3972-3976.
Offenberg JH; Naumova YY; Turpin BJ; Eisenreich SJ; Morandi MT; Stock T; Colome SD; Winer AM; Spektor DM; Zhang J;
Weisel CP. (2004). Chlordanes in the indoor and outdoor air of three U.S. cities. EnvironSci Technol 38: 2760-2768.
Oftedal B; Brunekreef B; Nystad W; Madsen C; Walker SE; Nafstad P. (2008). Residential outdoor air pollution and lung
function in schoolchildren. Epidemiology 19: 129-137.
Oftedal B; Nafstad P; Magnus P; Bjorkly S; Skrondal A. (2003). Traffic related air pollution and acute hospital admission for
respiratory diseases in Drammen, Norway 1995-2000. European journal of epidemiology 18: 671-675.
Ogulei D; Hopke PK; Wallace LA. (2006). Analysis of indoor particle size distributions in an occupied townhouse using positive
matrix factorization. IndoorAir 16: 204-215.
Oh SM; Chung KH. (2006). Identification of mammalian cell genotoxins in respirable diesel exhaust particles by bioassay-
directed chemical analysis. Toxicol Lett 161: 226-235.
Okin PM; Devereux RB; Howard BV; Fabsitz RR; Lee ET; Welty TK. (2000). Assessment of QT interval and QT dispersion for
prediction of all-cause and cardiovascular mortality in American Indians: The Strong Heart Study. Circulation 101: 61-
66.
December 2008
R-51
DRAFT-DO NOT QUOTE OR CITE
-------
Oldham MJ; Robinson RJ. (2007). Predicted tracheobronchial and pulmonary deposition in a murine asthma model. Anatomical
record 290: 1309-1314.
O'Leary DH; Polak JF; Kronmal RA; Kittner SJ; Bond MG; Wolfson SK, Jr.; Bommer W; Price TR; Gardin JM; Savage PJ.
(1992). Distribution and correlates of sonographically detected carotid artery disease in the Cardiovascular Health Study.
The CHS Collaborative Research Group. Stroke; a journal of cerebral circulation 23: 1752-1760.
O'Leary DH; Polak IF; Kronmal RA; Manolio TA; Burke GL; Wolfson SK, Jr. (1999). Carotid-artery intima and media thickness
as a risk factor for myocardial infarction and stroke in older adults. Cardiovascular Health Study Collaborative Research
Group. The New England journal of medicine 340: 14-22.
Olfert JS; Kulkarni P; Wang J. (2008). Measuring aerosol size distributions with the fast integrated mobility spectrometer. J
Aerosol Sci.
Oliveira A; Pampulha ME. (2006). Effects of long-term heavy metal contamination on soil microbial characteristics. Journal of
Bioscience and Bioengineering 102: 157-161.
Olivier JGJ; Bouwman AF; Berdowski JJM; Veldt C; Bloos JPJ; Visschedijk AJH; ZandveldPYJ; Haverlag JL. (1996).
Description of EDGAR Version 2.0: A set of global emission inventories of greenhouse gases and ozone-depleting
substances for all anthropogenic and most natural sources on a per country basis and on 1 degree x 1 degree grid.
Olivier JGJ; Bouwman AF; Berdowski JJM; Veldt C; Bloos JPJ; Visschedijk AJH; van derMaas CWM; ZandveldPYJ. (1999).
Sectoral emission inventories of greenhouse gases for 1990 on a per country basis as well as on l°x 1°. Environmental
Science and Policy 2: 241-263.
Olson DA; Norris GA. (2005). Sampling artifacts in measurement of elemental and organic carbon: Low-volume sampling in
indoor and outdoor environments. Atmospheric Environment 39: 5437-5445.
O'Neill MS; Hajat S; Zanobetti A; Ramirez-Aguilar M; Schwartz J. (2005b). Impact of control for air pollution and respiratory
epidemics on the estimated associations of temperature and daily mortality. Int J Biometeorol 50: 121-129.
O'Neill MS; Jerrett M; Kawachi I; Levy JI; Cohen AJ; Gouveia N; Wilkinson P; Fletcher T; Cifuentes L; Schwartz J. (2003).
Health, Wealth, and Air Pollution: Advancing Theory and Methods. Environmental health perspectives 111: 1861-1870.
O'Neill MS; Veves A; Sarnat JA; Zanobetti A; Gold DR; Economides PA; Horton ES; Schwartz J. (2007). Air pollution and
inflammation in type 2 diabetes: a mechanism for susceptibility. Occupational and environmental medicine 64: 373-379.
O'Neill MS; Veves A; Zanobetti A; Sarnat JA; Gold DR; Economides PA; Horton ES; Schwartz J. (2005a). Diabetes enhances
vulnerability to particulate air pollution-associated impairment in vascular reactivity and endothelial function. Circulation
111: 2913-2920.
Ono N; Oshio S; Niwata Y; Yoshida S; Tsukue N; Sugawara I; Takano H; Takeda K. (2007). Prenatal exposure to diesel exhaust
impairs mouse spermatogenesis. Inhalation toxicology 19: 275-281.
Oortgiesen M; Veronesi B; Eichenbaum G; Kiser PF; Simon SA. (2000). Residual oil fly ash and charged polymers activate
epithelial cells and nociceptive sensory neurons. American Journal of Physiology- Lung Cellular and Molecular
Physiology 278: 683-695.
Ormrod DP. (1984). Impact of trace element pollution on plants. In Treshow M (Ed.), Air pollution and plant life (pp. 291-319).
Chichester, United Kingdom: John Wiley & Sons.
Orsini DA; Ma Y; Sullivan A; Sierau B; Baumann K; Weber RJ. (2003). Refinements to the particle-into-liquid sampler (PILS)
for ground and airborne measurements of water soluble aerosol composition. Atmospheric Environment 37: 1243-1259.
Ostro B; Broadwin R; Green S; Feng WY; Lipsett M. (2006). Fine particulate air pollution and mortality in nine California
counties: results from CALFINE. Environ Health Perspect 114: 29-33.
Ostro B; Feng WY; Broadwin R; Green S; Lipsett M. (2007). The effects of components of fine particulate air pollution on
mortality in California: results from CALFINE. Environ Health Perspect 115: 13-19.
Ostro B; Feng WY; Broadwin R; Malig B; Green S; Lipsett M. (2008). The Impact of Components of Fine Particulate Matter on
Cardiovascular Mortality in Susceptible Subpopulations. Occupational and environmental medicine.
Otnyukova T. (2007). Epiphytic lichen growth abnormalities and element concentrations as early indicators of forest decline.
Environmental Pollution 146: 359-365.
Otvos E; Pazmandi T; Tuba Z. (2003). First national survey of atmospheric heavy metal deposition in Hungary by the analysis of
mosses. The Science of the total environment 309: 151-160.
Ozdilek HG; Mathisen PP; Pellegrino D. (2007). Distribution of heavy metals in vegetation surrounding the Blackstone River,
USA: Considerations regarding sediment contamination and long term metals transport in freshwater riverine ecosystems.
Journal of Environmental Biology 28: 493-502.
Ozkaynak H; Thurston GD. (1987). Associations between 1980 U.S. mortality rates and alternative measures of airborne particle
concentration. Risk Anal 7: 449-461.
Ozsahin K; Tugrul A; Mert S; Yuksel M; Tugrul G. (2006). Evaluation of pulmonary alveolo-capillary permeability in Type 2
diabetes mellitus Using technetium 99mTc-DTPA aerosol scintigraphy and carbon monoxide diffusion capacity. Journal
of Diabetes and Its Complications 20: 205-209.
Paatero P; Tapper U. (1994). Positive matrix factorization: A non-negative factor model with optimal utilization of error
estimates of data values. Environmetrics 5:111-126.
Pacheco AMG; Freitas MC. (2004). Are lower epiphytes really that better than higher plants for indicating airborne
contaminants? An insight into the elemental contents of lichen thalli and tree bark by INAA. Journal of Radioanalytical
and Nuclear Chemistry 259: 27-33.
December 2008
R-52
DRAFT-DO NOT QUOTE OR CITE
-------
Palli D; Saieva C; Munnia A; Peluso M; Grechi D; Zanna I; Caini S; Decarli A; Sera F; Masala G. (2008). DNA adducts and
PM(10) exposure in traffic-exposed workers and urban residents from the EPIC-Florence City study. The Science of the
total environment 403: 105-112.
Pandian MD; Behar JV; Ott WR; Wallace LA; Wilson AL; Colome SD; Koontz M. (1998). Correcting Errors in the Nationwide
Data Base of Residential Air Exchange Rates. Journal of exposure analysis and environmental epidemiology 8: 577-586.
Pandis SN. (2004). Atmospheric aerosol processes. In McMurry PH, Shepard, M., Vickery, J.S. (Ed.), Particulate Matter Science
for Policy Makers: A NARSTO Assessment. Cambridge, UK: Cambridge University Press.
Pang Y; Turpin BJ; Gundel LA. (2006). On the Importance of Organic Oxygen for Understanding Organic Aerosol Particles.
Aerosol Science and Technology 40: 128-133.
Park J; Mitchell MJ; McHale PJ; Christopher SF; Myers TP. (2003). Interactive effects of changing climate and atmospheric
deposition on N and S biogeochemistry in a forested watershed of the Adirondack Mountains, New York State. Global
Change Biol 9: 1602-1619.
Park JW; Lim YH; Kyung SY; An CH; Lee SP; Jeong SH; Ju YS. (2005a). Effects of ambient particulate matter on peak
expiratory flow rates and respiratory symptoms of asthmatics during Asian dust periods in Korea. Respirology 10: 470-
476.
Park K; Chow JC; Watson JG; Trimble DL; Doraiswamy P; Arnott WP; Stroud KR; Bowers K; Bode R. (2006a). Comparison of
Continuous and Filter-Based Carbon Measurements at the Fresno Supersite. Journal of the Air & Waste Management
Association 56.
Park RJ; Stenchikov GL; Pickering KE; Dickerson RR; Allen DJ; Kondragunta S. (2001). Regional air pollution and its radiative
forcing- Studies with a single-column chemical and radiation transport model. Journal of Geophysical Research D
Atmospheres 106: 28.
Park SK; O'Neill MS; Vokonas PS; Sparrow D; Schwartz J. (2005b). Effects of air pollution on heart rate variability: The VA
Normative Aging Study. Environ Health Perspect 113: 304-309.
Park SK; O'Neill MS; Vokonas PS; Sparrow D; Spiro A, 3rd; Tucker KL; Suh H; Hu H; Schwartz J. (2008b). Traffic-related
particles are associated with elevated homocysteine: the VA normative aging study. American journal of respiratory and
critical care medicine 178: 283-289.
Park SK; O'Neill MS; Vokonas PS; Sparrow D; Wright RO; Coull B; Nie H; Hu H; Schwartz J. (2008a). Air pollution and heart
rate variability: effect modification by chronic lead exposure. Epidemiology 19:111-120.
Park SK; O'Neill MS; Wright RO; Hu H; Vokonas PS; Sparrow D; Suh H; Schwartz J. (2006b). HFE Genotype, Particulate Air
Pollution, and Heart Rate Variability: A Gene-Environment Interaction. Circulation 114: 2798-2805.
Parker JD; Mendola P; Woodruff TJ. (2008). Preterm birth after the Utah valley steel mill closure: a natural experiment.
Epidemiology 19: 820-823.
Parker JD; Woodruff TJ. (2008). Influences of study design and location on the relationship between particulate matter air
pollution and birthweight. Paediatr Perinat Epidemiol 22: 214-227.
Parker JD; Woodruff TJ; Basu R; Schoendorf KC. (2005). Air pollution and birth weight among term infants in California.
Pediatrics 115: 121-128.
Parrish ZD; White JC; Isleyen M; Gent MPN; Iannucci-Berger W; Eitzer BD; Kelsey JW; Mattina MI. (2006). Accumulation of
weathered polycyclic aromatic hydrocarbons (PAHs) by plant and earthworm species. Chemosphere 64: 609-618.
Peacock JL; Symonds P; Jackson P; Bremner SA; Scarlett JF; Strachan DP; Anderson HR. (2003). Acute effects of winter air
pollution on respiratory function in schoolchildren in southern England. Occupational and environmental medicine 60:
82-89.
Pedersen M; Vinzents P; Petersen JH; Kleinjans JC; Plas G; Kirsch-Volders M; Dostal M; Rossner P; Beskid O; Sram RJ; Merlo
DF; Knudsen LE. (2006). Cytogenetic effects in children and mothers exposed to air pollution assessed by the frequency
of micronuclei and fluorescence in situ hybridization (FISH): a family pilot study in the Czech Republic. Mutat Res 608:
112-120.
Peel JL; Metzger KB; Klein M; Flanders WD; Mulholland JA; Tolbert PE. (2007). Ambient air pollution and cardiovascular
emergency department visits in potentially sensitive groups. Am J Epidemiol 165: 625-633.
Peel JL; Tolbert PE; Klein M; Metzger KB; Flanders WD; Todd K; Mulholland JA; Ryan PB; Frumkin H. (2005). Ambient air
pollution and respiratory emergency department visits. Epidemiology 16: 164-174.
Pekkanen J; Peters A; Hoek G; Tiittanen P; Brunekreef B; de Hartog J; Heinrich J; Ibald-Mulli A; Kreyling WG; Lanki T;
Timonen KL; Vanninen E. (2002). Particulate air pollution and risk of ST-segment depression during repeated
submaximal exercise tests among subjects with coronary heart disease: the Exposure and Risk Assessment for Fine and
Ultrafme Particles in Ambient Air (ULTRA) study. Circulation 106: 933-938.
Peled R; Friger M; Bolotin A; Bibi H; Epstein L; Pilpel D; Scharf S. (2005). Fine particles and meteorological conditions are
associated with lung function in children with asthma living near two power plants. Public Health 119: 418-425.
Pelletier B; Santer R; Vidot J. (2007). Retrieving of particulate matter from optical measurements: A semiparametric approach. J
Geophys Res 112.
Penard-Morand C; Charpin D; Raherison C; Kopferschmitt C; Caillaud D; Lavaud F; Annesi-Maesano I. (2005). Long-term
exposure to background air pollution related to respiratory and allergic health in schoolchildren. Clin Exp Allergy 35:
1279-1287.
Peng RD; Chang HH; Bell ML; McDermott A; Zeger SL; Samet JM; Dominici F. (2008). Coarse particulate matter air pollution
and hospital admissions for cardiovascular and respiratory diseases among Medicare patients. Jama 299: 2172-2179.
December 2008
R-53
DRAFT-DO NOT QUOTE OR CITE
-------
Peng RD; Dominici F; Pastor-Barriuso R; Zeger SL; Samet JM. (2005). Seasonal analyses of air pollution and mortality in 100
US cities. Am J Epidemiol 161: 585-594.
Penn A; Murphy G; Barker S; Henk W; Penn L. (2005). Combustion-derived ultrafme particles transport organic toxicants to
target respiratory cells. Environ Health Perspect 113: 956-963.
PennaMLF; Duchiade MP. (1991). Air pollution and infant mortality from pneumonia in the Rio de Janeiro metropolitan area.
Bull PAHO 25: 47-54.
Pennanen AS; Sillanpaa M; Hillamo R; Quass U; John AC; Branis M; Hunova I; Meliefste K; Janssen NA; Koskentalo T;
Castano-Vinyals G; Bouso L; Chalbot MC; Kavouras IG; Salonen RO. (2007). Performance of a high-volume cascade
impactor in six European urban environments: mass measurement and chemical characterization of size-segregated
particulate samples. The Science of the total environment 374: 297-310.
Pennanen T; Frostegard A; Fritze El; Baath E. (1996). Phospholipid fatty acid composition and heavy metal tolerance of soil
microbial communities along two heavy metal-polluted gradients in coniferous forests. Applied and Environmental
Microbiology 62: 420^128.
Penner JE, et al. (2001). Aerosols, their direct and indirect effects. In Eloughton JT, et al. (Ed.), Climate Change 2001: The
Scientific Basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on
Climate Change. Cambridge and New York: Cambridge University Press.
Penner JE. (2003). Comment on "Control of fossil-fuel particulate black carbon and organic matter, possibly the most effective
method of slowing global warming. J Geophys 108.
Penner JE; Zhang SY; Chuang CC. (2003). Soot and smoke aerosol may not warm climate. Journal of Geophysical Research
108(D21), 4657, doi:10.1029/2003JD003409.
Pennisi E. (2004). The secret life of fungi. Science 304: 1620-1622.
Penttinen P; Vallius M; Tiittanen P; Ruuskanen J; Pekkanen J. (2006). Source-specific fine particles in urban air and respiratory
function among adult asthmatics. Inhal Toxicol 18: 191-198.
Pereira CEL; Heck TG; Saldiva PHN; Rhoden CR. (2007). Ambient particulate air pollution from vehicles promotes lipid
peroxidation and inflammatory responses in rat lung. Brazilian journal of medical and biological research 40: 1353-1359.
Pereira LAA; Loomis D; Conceifao GMS; Braga ALF; Areas RM; Kishi HS; Singer JM; Bohm GM; Saldiva PHN. (1998).
Association between air pollution and intrauterine mortality in Sao Paulo, Brazil. Environ Health Perspect 106: 325-329.
Peretz A; Kaufman JD; Trenga CA; Allen J; Carlsten C; Aulet MR; Adar SD; Sullivan JH. (2008b). Effects of diesel exhaust
inhalation on heart rate variability in human volunteers. Environ Res 107: 178-184.
Peretz A; Peck EC; Bammler TK; Beyer RP; Sullivan JH; Trenga CA; Srinouanprachnah S; Farin FM; Kaufman JD. (2007).
Diesel exhaust inhalation and assessment of peripheral blood mononuclear cell gene transcription effects: an exploratory
study ofhealthy human volunteers. Inhal Toxicol 19: 1107-1119.
Peretz A; Sullivan JH; Leotta DF; Trenga CA; Sands FN; Allen J; Carlsten C; Wilkinson CW; Gill EA; Kaufman JD. (2008a).
Diesel exhaust inhalation elicits acute vasoconstriction in vivo. Environ Health Perspect 116: 937-942.
Perez L; Tobias A; Querol X; Kunzli N; Pey J; Alastuey A; Viana M; Valero N; Gonzalez-Cabre M; Sunyer J. (2008). Coarse
particles from Saharan dust and daily mortality. Epidemiology 19: 800-807.
Persson E; Henriksson J; Tallkvist J; Rouleau C; Tjalve H. (2003). Transport and subcellular distribution of intranasally
administered zinc in the olfactory system of rats and pikes. Toxicology 191: 97-108.
Petaja T. (2006). Detection Efficiency of a Water-Based TSI Condensation Particle Counter 3785. Aerosol Science and
Technology 40: 1090-1097.
Peters A; Dockery DW; Muller JE; Mittleman MA. (2001). Increased particulate air pollution and the triggering of myocardial
infarction. Circulation 103: 2810-2815.
Peters A; Liu E; Verrier RL; Schwartz J; Gold DR; Mittleman M; Baliff J; Oh JA; Allen G; Monahan K; Dockery DW. (2000).
Air pollution and incidence of cardiac arrhythmia. Epidemiology 11: 11-17.
Peters A; von Klot S; Heier M; Trentinaglia I; Cyrys J; Hormann A; Hauptmann M; Wichmann HE; Lowel H. (2005). Particulate
air pollution and nonfatal cardiac events. Part I. Air pollution, personal activities, and onset of myocardial infarction in a
case-crossover study. Res Rep Health Eff Inst: 1-66; discussion 67-82, 141-148.
Peters A; von Klot S; Heier M; Trentinaglia I; Hormann A; Wichmann HE; Lowel H. (2004). Exposure to traffic and the onset of
myocardial infarction. The New England journal of medicine 351: 1721-1730.
Peters AJ; Gregor DJ; Teixeira CF; Jones NP; Spencer C. (1995). The recent depositional trend of polycyclic aromatic
hydrocarbons and elemental carbon to the Agassiz Ice cap, Ellesmere Island, Canada. The Science of the total
environment 160/161: 167-179.
Peters JM; Avol E; Navidi W; London SJ; Gauderman WJ; Lurmann F; Linn WS; Margolis H; Rappaport E; Gong H; Thomas
DC. (1999). A study of twelve Southern California communities with differing levels and types of air pollution. I.
Prevalence of respiratory morbidity. American journal of respiratory and critical care medicine 159: 760-767.
Peters K; Eiden R. (1992). Modelling the dry deposition velocity of aerosol particles to a spruce forest. Atmospheric
Environment 26A: 2555-2564.
Peters TM. (2006). Use of the Aerodynamic Particle Sizer to Measure Ambient PM 10-2. 5: The Coarse Fraction of PM 10.
Journal of the Air & Waste Management Association 56.
Peterson BT; Dickerson KD; James HL; Miller EJ; McLarty JW; Holiday DB. (1989). Comparison of three tracers for detecting
lung epithelial injury in anesthetized sheep. Journal of applied physiology 66: 2374-2383.
December 2008
R-54
DRAFT-DO NOT QUOTE OR CITE
-------
Petrovic S; Urch B; Brook J; Datema J; Purdham J; Liu L; Lukic Z; Zimmerman B; Toiler G; Downar E; Corey P; Tarlo S;
Broder I; Dales R; Silverman F. (2000). Cardiorespiratory effects of concentrated ambient PM2.5: A pilot study using
controlled human exposures. Inhalation Toxicology 12: 173-188.
PettijohnFJ. (1957). Sedimentary rocks: Harper New York.
Pfirman SL; Kogeler JW; Rigor I. (1997). Potential for rapid transport of contaminants from the Kara Sea. Science of the Total
Environment, The 202: 111 -122.
Phalen RF; Oldham MJ. (2006). Aerosol dosimetry considerations. Clinics in occupational and environmental medicine 5: 773-
784.
Phalen RF; Oldham MJ; Mautz WJ. (1989). Aerosol Deposition in the Nose As a Function of Body Size. Health Physics 57: 299.
Phalen RF; Oldham MJ; Wolff RK. (2008). The relevance of animal models for aerosol studies. J Aerosol Med Pulm Drug Deliv
21: 113-124.
Phares DJ; Rhoads KP; Johnston MV; Wexler AS. (2003). Size-resolved ultrafme particle composition analysis 2. Houston.
Journal of Geophysical Research (Atmospheres) 108.
Phuleria HC; Geller MD; Fine PM; Sioutas C. (2006). Size-Resolved emissions of organic tracers from light-and heavy-duty
vehicles measured in a California roadway tunnel. Environ Sci Technol 40: 4109-4118.
Phuleria HC; Sheesley RJ; Schauer JJ; Fine PM; Sioutas C. (2007). Roadside measurements of size-segregated particulate
organic compounds near gasoline and diesel-dominated freeways in Los Angeles, CA. Atmospheric Environment 41:
4653-4671.
Piedrahita JA; Zhang SH; Hagaman JR; Oliver PM; Maeda N. (1992). Generation of mice carrying a mutant apolipoprotein E
gene inactivated by gene targeting in embryonic stem cells. Proc Natl Acad Sci U S A 89: 4471-4475.
Pierse N; Rushton L; Harris RS; Kuehni CE; Silverman M; Grigg J. (2006). Locally generated particulate pollution and
respiratory symptoms in young children. Thorax 61: 216-220.
Pietropaoli AP; Frampton MW; Hyde RW; Morrow PE; Oberdorster G; Cox C; Speers DM; Frasier LM; Chalupa DC; Huang
LS; Utell MJ. (2004). Pulmonary function, diffusing capacity, and inflammation in healthy and asthmatic subjects
exposed to ultrafme particles. Inhalation Toxicology 16: 59-72.
Pignoli P; Tremoli E; Poli A; Oreste P; Paoletti R. (1986). Intimal plus medial thickness of the arterial wall: a direct measurement
with ultrasound imaging. Circulation 74: 1399-1406.
Pigorini F; Maini CL; Pau F; Giosue S. (1988). The influence of age on the pulmonary clearance of 99Tcm-DTPA radioaerosol.
Nucl Med Commun 9: 965-971.
Pinkerton KE; Zhou Y-M; Teague SV; Peake JL; Walther RC; Kennedy IM; Leppert VJ; Aust AE. (2004). Reduced lung cell
proliferation following short-term exposure to ultrafme soot and iron particles in neonatal rats: key to impaired lung
growth? Inhalation toxicology 16: 73-81.
Piol MN; Lopez AG; Mino LA; Dos Santos Afonso M; Verrengia Guerrero NR. (2006). The impact of particle-bound cadmium
on bioavailability and bioaccumulation: A pragmatic approach. Environmental Science and Technology 40: 6341-6347.
Pires-Neto RC; Lichtenfels AJ; Soares SR; Macchione M; Saldiva PHN; Dolhnikoff M. (2006). Effects of Sao Paulo air pollution
on the upper airways of mice. Environmental research 101: 356-361.
Pirjola L; Parviainen H; Hussein T; Valli A; Hameri K; Aaalto P; Virtanen A; Keskinen J; Pakkanen TA; Makela T. (2004).
"Sniffer"—a novel tool for chasing vehicles and measuring traffic pollutants. Atmospheric Environment 38: 3625-3635.
Pitchford M; Maim W; Schichtel B; Kumar N; Lowenthal D; Hand J. (2007). Revised algorithm for estimating light extinction
from IMPROVE particle speciation data. J Air Waste Manag Assoc 57: 1326-1336.
Pitchford ML; Malm WC. (1994). Development and applications of a standard visual index. Atmospheric environment(1994) 28:
1049-1054.
Pitchford ML; Polkowsky BV; McGown MR; Malm WC; Molenar JV; Mauch L. (1990). Percent change in extinction
coefficient: A proposed approach for federal visibility protection strategy.
Pitchford ML; Schichtel BA; Gebhart KA; Barna MG; Malm WC; Tombach IH; Knipping EM. (2005). Reconciliation and
Interpretation of the Big Bend National Park Light Extinction Source Apportionment: Results from the Big Bend
Regional Aerosol and Visibility Observational Study- Part II. Journal of the Air & Waste Management Association 55.
Pleis JR; Lethbridge-Cejku M. (2007). Summary health statistics for US adults: National Health Interview Survey, 2006. Vital
and health statistics Series 10, Data from the National Health Survey: 1.
Polidori A; Arhami M; Sioutas C; Delfino RJ; Allen R. (2007). Indoor/Outdoor relationships, trends, and carbonaceous content
of fine particulate matter in retirement homes of the Los Angeles Basin. J Air Waste ManagAssoc 57: 366-379.
Polidori A; Turpin B; Meng QY; Lee JH; Weisel C; Morandi M; Colome S; Stock T; Winer A; Zhang J; Kwon J; Alimokhtari S;
Shendell D; Jones J; Farrar C; Maberti S. (2006). Fine organic particulate matter dominates indoor-generated PM2.5 in
RIOPA homes. J ExpoSci EnvironEpidemiol 16: 321-331.
Poma A; Limongi T; Pisani C; Granato V; Picozzi P. (2006). Genotoxicity induced by fine urban air particulate matter in the
macrophages cell line RAW 264.7. Toxicol In Vitro 20: 1023-1029.
Pope CA, 3rd. (1989). Respiratory disease associated with community air pollution and a steel mill, Utah Valley. Am J Public
Health 79: 623-628.
Pope CA, 3rd; Burnett RT. (2007). Confounding in air pollution epidemiology: the broader context. Epidemiology 18: 424-426;
discussion 427-428.
Pope CA, 3rd; Burnett RT; Thun MJ; Calle EE; Krewski D; Ito K; Thurston GD. (2002). Lung cancer, cardiopulmonary
mortality, and long-term exposure to fine particulate air pollution. JAMA 287: 1132-1141.
December 2008
R-55
DRAFT-DO NOT QUOTE OR CITE
-------
Pope CA, 3rd; Burnett RT; Thurston GD; Thun MJ; Calle EE; Krewski D; Godleski JJ. (2004b). Cardiovascular mortality and
long-term exposure to particulate air pollution: epidemiological evidence of general pathophysiological pathways of
disease. Circulation 109: 71-77.
Pope CA, 3rd; Muhlestein IB; May HT; Renlund DG; Anderson JL; Home BD. (2006). Ischemic heart disease events triggered
by short-term exposure to fine particulate air pollution. Circulation 114: 2443-2448.
Pope CA, 3rd; Rodermund DL; Gee MM. (2007). Mortality effects of a copper smelter strike and reduced ambient sulfate
particulate matter air pollution. Environ Health Perspect 115: 679-683.
Pope CA, 3rd; Thun MJ; Namboodiri MM; Dockery DW; Evans JS; Speizer FE; Heath CW, Jr. (1995). Particulate air pollution
as a predictor of mortality in a prospective study of U.S. adults. American journal of respiratory and critical care medicine
151: 669-674.
Pope CA; Dockery DW. (2006). Health effects of fine particulate air pollution: Lines that connect. J Air Waste Manag Assoc 56:
709-742.
Pope CA; Hansen ML; Long RW; Nielsen KR; Eatough NL; Wilson WE; Eatough DJ. (2004a). Ambient particulate air
pollution, heart rate variability, and blood markers of inflammation in a panel of elderly subjects. Environ Health Perspect
112: 339-345.
Poschl U. (2005). Atmospheric aerosols: composition, transformation, climate and health effects. Angew Chem Int Ed 44: 7520-
7540.
Pouliot G; Pace TG; Roy B; Pierce T; Mobley D. (2008). Development of a biomass burning emissions inventory by combining
satellite and ground-based information. Journal of Applied Remote Sensing 2: 021501-021517.
Pourazar J; Blomberg A; Kelly FJ; Davies DE; Wilson SJ; Holgate ST; Sandstrom T. (2008). Diesel exhaust increases EGFR and
phosphorylated C-terminal Tyr 1173 in the bronchial epithelium. Part Fibre Toxicol 5: 8.
Pourazar J; Mudway IS; Samet JM; Helleday R; Blomberg A; Wilson SJ; Frew AJ; Kelly FJ; Sandstrom T. (2005). Diesel
exhaust activates redox-sensitive transcription factors and kinases in human airways. Am J Physiol Lung Cell Mol
Physiol 289: L724-730.
Prasad MNV; De Oliveira Freitas HM. (2003). Metal hyperaccumulation in plants - Biodiversity prospecting for. Electronic
Journal of Biotechnology 6: 110-146.
Preston CM; Schmidt MWI. (2006). Black (pyrogenic) carbon: a synthesis of current knowledge and uncertainties with special
consideration of boreal regions. Biogeosciences 3: 397-420.
Preutthipan A; Udomsubpayakul U; Chaisupamongkollarp T; Pentamwa P. (2004). Effect of PM10 pollution in Bangkok on
children with and without asthma. Pediatr Pulmonol 37: 187-192.
Price M; Bulpitt S; Meyer MB. (2003). A comparison of PM10 monitors at a Kerbside site in the northeast of England.
Atmospheric Environment 37: 4425-4434.
Pritchett LC; Cooper JA. (1985). Aerosol Characterization Study of Anchorage, Alaska: Chemical Analysis and Source
Apportionment. Prepared for Municipality of Anchorage Air Pollution Control Agency, Anchorage, AK, by NEA
Laboratories, Beaverton, OR.
Prospero JM. (1996). Saharan dust transport over the North Atlantic Ocean and Mediterranean: an overview. In Guerzoni S,
Chester R (Eds.), The Impact of Desert Dust Across the Mediterranean (pp. 133-152). The Netherlands: Kluwer
Academic Publishers.
Pryor SC. (1996). Assessing public perception of visibility for standard setting exercises. Atmospheric Environment 30: 2705-
2716.
Pryor SC; Binkowski FS. (2004). An analysis of the time scales associated with aerosol processes during dry deposition. Aerosol
Science and Technology 38: 1091-1098.
Prystowsky EN; Benson DW, Jr.; Fuster V; Hart RG; Kay GN; Myerburg RJ; Naccarelli GV; Wyse DG. (1996). Management of
patients with atrial fibrillation. A Statement for Healthcare Professionals. From the Subcommittee on Electrocardiography
and Electrophysiology, American Heart Association. Circulation 93: 1262-1277.
Puchelle E; Zahm JM; Bertrand A. (1979). Influence of age on bronchial mucociliary transport. Scand J Respir Dis 60: 307-313.
Puett RC; Schwartz J; Hart JE; Yanosky JD; Speizer FE; Suh H; Paciorek CJ; Neas LM; Laden F. (2008). Chronic Particulate
Exposure, Mortality, and Coronary Heart Disease in the Nurses' Health Study. Am J Epidemiol.
Putaud JP; Raes F; Van Dingenen R; Briiggeman E; Facchini MC; Decesari S; Fuzzi S; Gehrig R; Hiiglin C; Laj P; Lorbeer G;
Maenhaut W; Mihalopoulos N; Miiller K; Querol X; Rodriguez S; Schneirder J; Spindler G; Brink H; Torseth K;
Wiedensohler A. (2004). A European aerosol phenomenology-2: chemical characteristics of particulate matter at
kerbside, urban, rural and background sites in Europe. Atmospheric Environment 38: 2579-2595.
Qin X; Prather KA. (2006). Impact of biomass emissions on particle chemistry during the California Regional Particulate Air
Quality Study. International Journal of Mass Spectrometry 258: 142-150.
Quante M. (2004). The role of clouds in the climate system. Paper presented at the Journal De Physique. IV : JP.
Quinn PK; Bates TS; Baynard T; Clarke AD; Onasch TB; Wang W; Rood MJ; Andrews E; Allan J; Carrico CM; Coffman D;
Worsnop D. (2005). Impact of particulate organic matter on the relative humidity dependence of light scattering: A
simplified parameterization. Geophysical Research Letters 32: 1-4.
Raabe OG; Al-Bavati MA; Teague SV; Rasolt A. (1985). Regional deposition of inhaled monodisperse coarse and fine aerosol
particles in small laboratory animals. PB-86-116555/XAB, California Univ., Davis (USA),
Rabinovitch N; Strand M; Gelfand EW. (2006). Particulate levels are associated with early asthma worsening in children with
persistent disease. American journal of respiratory and critical care medicine 173: 1098-1105.
December 2008
R-56
DRAFT-DO NOT QUOTE OR CITE
-------
Rabinovitch N; Zhang L; Murphy JR; Vedal S; Dutton SJ; Gelfand EW. (2004). Effects of wintertime ambient air pollutants on
asthma exacerbations in urban minority children with moderate to severe disease. J Allergy Clin Immunol 114: 1131-
1137.
Radke LF; Coakley JA; King MD. (1989). Direct and remote sensing observations of the effects of ships on clouds. Science 246:
1146-1149.
Raizenne M; Neas LM; Damokosh AI; Dockery DW; Spengler JD; Koutrakis P; Ware JH; Speizer FE. (1996). Elealth effects of
acid aerosols on North American children: pulmonary function. Environ Elealth Perspect 104: 506-514.
Rajapaksha R; M. T-K; Baath E. (2004). Metal toxicity affects fungal and bacterial activities in soil differently. Applied and
Environmental Microbiology 5: 2966 - 2973.
Ramage L; Guy K. (2004). Expression of C-reactive protein and heat-shock protein-70 in the lung epithelial cell line A549, in
response to PM10 exposure. Inhal Toxicol 16: 447-452.
Ramanathan V; Crutzen PJ; Kiehl JT; Rosenfeld D. (2001). Aerosols, climate, and the hydrological cycle. Science 294: 2119-
2124.
Ramaswamy V; Boucher O; Elaigh J; Elauglustaine D; Elaywood J; Myhre G; Nakajima T; Shi G; Solomon S; Betts RE. (2001).
Radiative Forcing of Climate Change: PNNL-SA-39648, Eloughton, JT et al; Cambridge University Press, New York,
NY, United States (US).
Ramos L; Elernandez LM; Gonzalez MJ. (1994). Sequential fractionation of copper, lead, cadmium, and zinc in soils from or near
Donana National Park. J Environ Qual 23: 50-57.
Randerson JT; Liu H; Flanner MG; Chambers SD; Jin Y; Eless PG; Pfister G; Mack MC; Treseder KK; Welp LR; Chapin FS;
Elarden JW; Goulden ML; Lyons E; Neff JC; Schuur EAG; Zender CS. (2006). The impact of boreal forest fire on
climate warming. Science 314: 1130-1132.
Rannug U; Sundvall A; Westerholm R; AlsbergT; Stenberg U. (1983). Some aspects of mutagenicity testing of the particulate
phase and the gas phase of diluted and undiluted automobile exhaust. ENVIRON SCI RES 1983.
Ranzi A; Gambini M; Spattini A; Galassi C; Sesti D; Bedeschi M; Messori A; Baroni A; Cavagni G; Lauriola P. (2004). Air
pollution and respiratory status in asthmatic children: hints for a locally based preventive strategy. AIRE study. European
journal of epidemiology 19: 567-576.
Rapport DJ; Whitford WG. (1999). How ecosystems respond to stress: common properties of arid and aquatic systems.
Bioscience 49: 193-203.
Rattigan OV; Hogrefe O; Felton HD; Schwab JJ; Roychowdhury UK; Husain L; Dutkiewicz VA; Demeijian KL. (2006). Multi-
year urban and rural semi-continuous PM2. 5 sulfate and nitrate measurements in New York state: Evaluation and
comparison with filter based measurements. Atmospheric Environment 40: 192-205.
RauschN; Nieminen T; Ukonmaanaho L; Roux GL; KrachlerM; Cheburkin AK; Bonani G; Shotyk W. (2005). Comparison of
atmospheric deposition of copper, nickel, cobalt, zinc, and cadmium recorded by Finnish peat cores with monitoring data
and emission records. Environmental Science and Technology 39: 5989-5998.
Reddy MS; Boucher O. (2007). Climate impact of black carbon emitted from energy consumption in the world's regions.
Geophysical Research Letters 34: 1802.
Reed MD; Barrett EG; Campen MJ; Divine KK; Gigliotti AP; McDonald JD; Seagrave JC; Mauderly JL; Seilkop SK; Swenberg
JA. (2008). Health Effects of Subchronic Inhalation Exposure to Gasoline Engine Exhaust. Inhal Toxicol: 1.
Reed MD; Campen MJ; Gigliotti AP; Harrod KS; McDonald JD; Seagrave JC; Mauderly JL; Seilkop SK. (2006). Health effects
of subchronic exposure to environmental levels of hardwood smoke. Inhalation Toxicology 18: 523-539.
Reed MD; Gigliotti AP; McDonald JD; Seagrave JC; Seilkop SK; Mauderly JL. (2004). Health effects of subchronic exposure to
environmental levels of diesel exhaust. Inhalation Toxicology 16: 177-193.
Rees SL; Robinson AL; Khlystov A; Stanier CO; Pandis SN. (2004). Mass balance closure and the Federal Reference Method for
PM2. 5 in Pittsburgh, Pennsylvania. Atmospheric Environment 38: 3305-3318.
Reff A; Weisel CP; Zhang J; Morandi M; Stock T; Colome S; Winer A; Turpin BJ; Offenberg JH. (2007). A functional group
characterization of organic PM2.5 exposure: Results from the RIOPA study. Atmospheric Environment 41: 4585-4598.
Regoli F; Gorbi S; Fattorini D; Tedesco S; Notti A; Machella N; Bocchetti R; Benedetti M; Piva F. (2006). Use of the land snail
Helix aspersa sentinel organism for monitoring ecotoxicologic effects of urban pollution: An integrated approach.
Environmental Health Perspectives 114: 63-69.
Reinfelder JR; Fisher NS; Luoma SN; Nichols JW; Wang W-X. (1998). Trace element trophic transfer in aquatic organisms: A
critique of the kinetic model approach. Science of the Total Environment 219: 117-135.
Reisinger LM. (1990). Analysis of airborne particles sampled in the southern Appalachian Mountains. Water Air and Soil
Pollution 50: 149-162.
Rengasamy A; Barger MW; Kane E; Ma JKH; Castranova V; Ma JYC. (2003). Diesel exhaust particle-induced alterations of
pulmonary phase I and phase II enzymes of rats. Journal of toxicology and environmental health Part A 66: 153-167.
Reponen T; Grinshpun SA; Trakumas S; Martuzevicius D; Wang ZM; LeMasters G; Lockey JE; Biswas P. (2003). Concentration
gradient patterns of aerosol particles near interstate highways in the Greater Cincinnati airshed. J Environ Monit 5: 557-
562.
Resnick HE; Lindsay RS; McDermott MM; Devereux RB; Jones KL; Fabsitz RR; Howard BV. (2004). Relationship of high and
low ankle brachial index to all-cause and cardiovascular disease mortality: the Strong Heart Study. Circulation 109: 733-
739.
December 2008
R-57
DRAFT-DO NOT QUOTE OR CITE
-------
Rhoden CR; Lawrence J; Godleski JJ; Gonzalez-Flecha B. (2004). N-acetylcysteine prevents lung inflammation after short-term
inhalation exposure to concentrated ambient particles. Toxicological sciences 79: 296-303.
Rhoden CR; Wellenius GA; Ghelfi E; Lawrence J; Gonzalez-Flecha B. (2005). PM-induced cardiac oxidative stress and
dysfunction are mediated by autonomic stimulation. Biochimica et Biophysica Acta 1725: 305-313.
Rice J. (2004). Comparison of Integrated Filter and Automated Carbon Aerosol Measurements at Research Triangle Park, North
Carolina. Aerosol Science and Technology 38: 23-36.
Rich DQ; Freudenberger RS; Ohman-Strickland P; Cho Y; Kipen HM. (2008). Right heart pressure increases after acute
increases in ambient particulate concentration. Environ Health Perspect 116: 1167-1171.
Rich DQ; Kim MH; Turner JR; Mittleman MA; Schwartz J; Catalano PJ; Dockery DW. (2006b). Association of ventricular
arrhythmias detected by implantable cardioverter defibrillator and ambient air pollutants in the St Louis, Missouri
metropolitan area. Occupational and environmental medicine 63: 591-596.
Rich DQ; Mittleman MA; Link MS; Schwartz J; Luttmann-Gibson H; Catalano PJ; Speizer FE; Gold DR; Dockery DW. (2006a).
Increased risk of paroxysmal atrial fibrillation episodes associated with acute increases in ambient air pollution. Environ
Health Perspect 114: 120-123.
Rich DQ; Schwartz J; Mittleman MA; Link M; Luttmann-Gibson H; Catalano PJ; Speizer FE; Dockery DW. (2005). Association
of short-term ambient air pollution concentrations and ventricular arrhythmias. Am J Epidemiol 161: 1123-1132.
Rich KE; Petkau J; Vedal S; Brauer M. (2004). A case-crossover analysis of particulate air pollution and cardiac arrhythmia in
patients with implantable cardioverter defibrillators. Inhal Toxicol 16: 363-372.
Riddle SG; Jakober CA; Robert MA; Cahill TM; Charles MJ; Kleeman MJ. (2007). Large PAHs detected in fine particulate
matter emitted from light-duty gasoline vehicles. Atmospheric Environment 41: 8658-8668.
Riediker M; Cascio WE; Griggs TR; Herbst MC; Bromberg PA; Neas L; Williams RW; Devlin RB. (2004b). Particulate matter
exposure in cars is associated with cardiovascular effects in healthy young men. American journal of respiratory and
critical care medicine 169: 934-940.
Riediker M; Devlin RB; Griggs TR; Herbst MC; Bromberg PA; Williams RW; Cascio WE. (2004a). Cardiovascular effects in
partrol officers are associated with fine particulate matter from brake wear and engine emissions. Particle Fibre Toxicol 1.
Rigby M; Toumi R. (2008). London air pollution climatology: Indirect evidence for urban boundary layer height and wind speed
enhancement. Atmospheric Environment 42: 4932-4947.
Riojas-Rodriguez H; Escamilla-Cejudo JA; Gonzalez-Hermosillo JA; Tellez-Rojo MM; Vallejo M; Santos-Burgoa C; Rojas-
Bracho L. (2006). Personal PM2.5 and CO exposures and heart rate variability in subjects with known ischemic heart
disease in Mexico City. Journal of Exposure Science and Environmental Epidemiology 16: 131-137.
Risom L; Mrller P; Loft S. (2005). Oxidative stress-induced DNA damage by particulate air pollution. Mutation Research-
Fundamental and Molecular Mechanisms of Mutagenesis 592: 119-137.
Ritz B; Wilhelm M. (2008). Ambient air pollution and adverse birth outcomes: methodologic issues in an emerging field. Basic
& Clinical Toxicology 102: 182-190.
Ritz B; Wilhelm M; Hoggatt KJ; Ghosh JK. (2007). Ambient air pollution and preterm birth in the environment and pregnancy
outcomes study at the University of California, Los Angeles. Am J Epidemiol 166: 1045-1052.
Ritz B; Wilhelm M; Zhao Y. (2006). Air pollution and infant death in southern California, 1989-2000. Pediatrics 118: 493-502.
Ritz B; Yu F; Chapa G; Fruin S. (2000). Effect of air pollution on preterm birth among children born in Southern California
between 1989 and 1993. Epidemiology 11: 502-511.
Ritz B; Yu F; Fruin S; Chapa G; Shaw GM; Harris JA. (2002). Ambient air pollution and risk of birth defects in Southern
California. Am J Epidemiol 155: 17-25.
Rivero DH; Soares SR; Lorenzi-Filho G; Saiki M; Godleski JJ; Antonangelo L; Dolhnikoff M; Saldiva PH. (2005a). Acute
cardiopulmonary alterations induced by fine particulate matter of Sao Paulo, Brazil. Toxicol Sci 85: 898-905.
Rivero DHRF; Sassaki C; Lorenzi-Filho G; Saldiva PHN. (2005b). PM2.5 induces acute electrocardiographic alterations in
healthy rats. Environmental Research 99: 262-266.
Roache PJ. (1998). Verification and validation in computational science and engineering. Albuquerque, NM Hermosa Publishers.
Roberts DL; Jones A. (2004). Climate sensitivity to black carbon aerosol from fossil fuel combustion. Journal of Geophysical
Research 109: 6202.
Roberts ES; Richards JH; Jaskot R; Dreher KL. (2003). Oxidative stress mediates air pollution particle-induced acute lung injury
and molecular pathology. Inhalation toxicology 15: 1327-1346.
Roberts JR; Young S-H; Castranova V; Antonini JM. (2007). Soluble metals in residual oil fly ash alter innate and adaptive
pulmonary immune responses to bacterial infection in rats. Toxicology and applied pharmacology 221: 306-319.
Robinson AL; Donahue NM; Rogge WF. (2006). Photochemical oxidation and changes in molecular composition of organic
aerosol in the regional context. J Geophy s Res 111.
Robinson AL; Donahue NM; Shrivastava MK; Weitkamp EA; Sage AM; Grieshop AP; Lane TE; Pierce JR; Pandis SN. (2007a).
Rethinking organic aerosols: semivolatile emissions and photochemical aging. Science 315: 1259-1262.
Robinson DL; Monro JM; Campbell EA. (2007b). Spatial variability and population exposure to PM2.5 pollution from
woodsmoke in a New South Wales country town. Atmospheric Environment 41: 5464-5478.
Rocher V; Azimi S; Gasperi J; Beuvin L; Muller M; Moilleron R; Chebbo G. (2004). Hydrocarbons and metals in atmospheric
deposition and roof runoff in central Paris. Water Air and Soil Pollution 159: 67-86.
December 2008
R-58
DRAFT-DO NOT QUOTE OR CITE
-------
Rodriguez C; Tonkin R; Heyworth J; Kusel M; De Klerk N; Sly PD; Franklin P; Runnion T; Blockley A; Landau L; Hinwood
AL. (2007). The relationship between outdoor air quality and respiratory symptoms in young children. Int J Environ
Health Res 17: 351-360.
Roeckner E; Stier P; Feichter J; Kloster S; Esch M; Fischer-Brans L. (2006). Impact of carbonaceous aerosol emissions on
regional climate change. Climate Dynamics 27: 553-571.
Rogers JF; Dunlop AL. (2006). Air pollution and very low birth weight infants: a target population? Pediatrics 118: 156-164.
Rojas-Bracho L; Suh HH; Catalano PJ; Koutrakis P. (2004). Personal exposures to particles and their relationships with personal
activities for chronic obstructive pulmonary disease patients living in Boston. J Air Waste ManagAssoc 54: 207-217.
Rojas-Martinez R; Perez-Padilla R; Olaiz-Fernandez G; Mendoza-Alvarado L; Moreno-Macias H; Fortoul T; McDonnell W;
Loomis D; Romieu I. (2007). Lung function growth in children with long-term exposure to air pollutants in Mexico City.
American journal of respiratory and critical care medicine 176: 377-384.
Roman HA; Walker KD; Walsh TL; Conner L; Richmond HM; Hubbell BJ; Kinney PL. (2008). Expert judgment assessment of
the mortality impact of changes in ambient fine particulate matter in the U.S. Environ Sci Technol 42: 2268-2274.
Romic M; Romic D. (2003). Heavy metals distribution in agricultural topsoils in urban area. Environmental Geology 43: 795-
805.
Romieu I; Garcia-Esteban R; Sunyer J; Rios C; Alcaraz-Zubeldia M; Velasco SR; Holguin F. (2008). The effect of
supplementation with omega-3 polyunsaturated fatty acids on markers of oxidative stress in elderly exposed to PM(2.5).
Environ Health Perspect 116: 1237-1242.
Romieu I; Ramirez-Aguilar M; Moreno-Macias H; Barraza-Villarreal A; Miller P; Hernandez-Cadena L; Carbajal-Arroyo LA;
Hernandez-Avila M. (2004). Infant mortality and air pollution: modifying effect by social class. J Occup Environ Med
46: 1210-1216.
Romieu I; Tellez-Rojo MM; Lazo M; Manzano-Patino A; Cortez-Lugo M; Julien P; Belanger MC; Hernandez-Avila M; Holguin
F. (2005). Omega-3 Fatty Acid Prevents Heart Rate Variability Reductions Associated with Particulate Matter. Am J
Respir Crit Care Med 172: 1534-1540.
Roosli M; Braun-Fahrlander C; Kunzli N; Oglesby L; Theis G; Camenzind M; Mathys P; Staehelin J. (2000). Spatial variability
of different fractions of particulate matter within an urban environment and between urban and rural sites. J Air Waste
ManagAssoc 50: 1115-1124.
Roosli M; Kunzli N; Braun-Fahrlander C; Egger M. (2005). Years of life lost attributable to air pollution in Switzerland: dynamic
exposure-response model. Int J Epidemiol 34: 1029-1035.
Roosli M; Theis G; Kunzli N; Staehelin J; Mathys P; Oglesby L; Camenzind M; Braun-Fahrlander C. (2001). Temporal and
spatial variation of the chemical composition of PM10 at urban and rural sites in the Basel area, Switzerland.
Atmospheric Environment 35: 3701-3713.
Rosas Perez I; Serrano J; Alfaro-Moreno E; Baumgardner D; Garcia-Cuellar C; Martin Del Campo JM; Raga GB; Castillejos M;
Colin RD; Osornio Vargas AR. (2007). Relations between PM10 composition and cell toxicity: a multivariate and
graphical approach. Chemosphere 67: 1218-1228.
Rosenfeld D. (1999). TRMM observed first direct evidence of smoke from forest fires inhibiting rainfall. Geophysical Research
Letters 26: 3105-3108.
Rosenfeld D. (2000). Suppression of rain and snow by urban and industrial air pollution. Science 287: 1793-1796.
Rosenfeld D; Dai J; Yu X; Yao Z; Xu X; Yang X; Du C. (2007). Inverse relations between amounts of air pollution and
orographic precipitation. Science 315: 1396-1398.
Rosenfeld D; Givati A. (2006). Evidence of orographic precipitation suppression by air pollution-induced aerosols in the western
United States. Journal of Applied Meteorology and Climatology 45: 893-911.
Rosenfeld D; Lahav R; Khain A; Pinsky M. (2002). The role of sea spray in cleansing air pollution over ocean via cloud
processes. Science 297: 1667-1670.
Rosenlund M; Berglind N; Pershagen G; Hallqvist J; Jonson T; Bellander T. (2006). Long-term exposure to urban air pollution
and myocardial infarction. Epidemiology 17: 383-390.
Rosenthal FS; Carney JP; dinger ML. (2008). Out-of-hospital cardiac arrest and airborne fine particulate matter: a case-
crossover analysis of emergency medical services data in Indianapolis, Indiana. Environ Health Perspect 116: 631-636.
Ross R. (1999). Atherosclerosis—an inflammatory disease. The New England journal of medicine 340: 115-126.
Rosselli W; Keller C; Boschi K. (2003). Phytoextraction capacity of trees growing on a metal contaminated soil. Plant and Soil
256: 265-272.
Rossner P; Svecova V; Milcova A; Lnenickova Z; Solansky N; Sram RJ. (2008). Seasonal variability of oxidative stress markers
in city bus drivers - Part I. Oxidative damage to DNA. Mutation Research-Fundamental and Molecular Mechanisms of
Mutagenesis 642: 14-20.
Roth C; Scheuch G; Stahlhofen W. (1993). Clearance of the human lungs for ultrafme particles. Journal of Aerosol SciencefJ
AEROSOL SCI] 24.
Roth CM; Goss K; Schwarzenbach RP. (2004). Sorption of diverse organic vapors to snow. Environmental Science and
Technology 38: 4078-4084.
Rothman KJ; Greenland S. (1998). Modern epidemiology (2 ed.). Philadelphia, PA: Lippincott-Raven Publishers.
Roubicek DA; Gutierrez-Castillo ME; Sordo M; Cebrian-Garcia ME; Ostrosky-Wegman P. (2007). Micronuclei induced by
airborne particulate matter from Mexico City. Mutat Res 631: 9-15.
December 2008
R-59
DRAFT-DO NOT QUOTE OR CITE
-------
Rouse RL; Murphy G; Boudreaux MJ; Paulsen DB; Penn AL. (2008). Soot nanoparticles promote biotransformation, oxidative
stress, and inflammation in murine lungs. Am J Respir Cell Mol Biol 39: 198-207.
Routledge HC; Manney S; Harrison RM; Ayres JG; Townend JN. (2006). Effect of inhaled sulphur dioxide and carbon particles
on heart rate variability and markers of inflammation and coagulation in human subjects. Heart 92: 220-227.
Rubes J; Selevan SG; Evenson DP; Zudova D; Vozdova M; Zudova Z; Robbins WA; Perreault SD. (2005). Episodic air pollution
is associated with increased DNA fragmentation in human sperm without other changes in semen quality. Hum Reprod
20: 2776-2783.
Ruckerl R; Greven S; Ljungman P; Aalto P; Antoniades C; Bellander T; Berglind N; Chrysohoou C; Forastiere F; Jacquemin B;
von Klot S; Koenig W; Kuchenhoff H; Lanki T; Pekkanen J; Perucci CA; Schneider A; Sunyer J; Peters A. (2007b). Air
pollution and inflammation (interleukin-6, C-reactive protein, fibrinogen) in myocardial infarction survivors. Environ
Health Perspect 115: 1072-1080.
Ruckerl R; Ibald-Mulli A; Koenig W; Schneider A; Woelke G; Cyrys J; Heinrich J; Marder V; Frampton M; Wichmann HE;
Peters A. (2006). Air pollution and markers of inflammation and coagulation in patients with coronary heart disease.
American journal of respiratory and critical care medicine 173: 432-441.
Ruckerl R; Phipps RP; Schneider A; Frampton M; Cyrys J; Oberdorster G; Wichmann HE; Peters A. (2007a). Ultrafine particles
and platelet activation in patients with coronary heart disease—results from a prospective panel study. Part Fibre Toxicol
4: 1.
Rudich Y; Donahue NM; Mentel TF. (2007). Aging of Organic Aerosol: Bridging the Gap Between Laboratory and Field
Studies. Annual Review of Physical Chemistry 58: 321.
Ruggenenti P; Remuzzi G. (2006). Time to abandon microalbuminuria? Kidney Int 70: 1214-1222.
Rundell KW, J.R. Hoffman, R. Caviston, R. Bulbulian, A.M. Hollenbach. (2007). Inhalation of ultrafine and fine particulate
matter disrupts systemic vascular function. Inhalation Toxicology 19: 133-140.
Russell A; Dermis R. (2000). NARSTO critical review of photochemical models and modeling. Atmospheric Environment 34:
2283-2324.
Russell M; Allen D; Collins D; Fraser M. (2004). Daily, Seasonal, and Spatial Trends in PM2. 5 Mass and Composition in
Southeast Texas. Aerosol Science and Technology 38: 14-26.
Rusu A-M; Jones GC; Chimonides PDJ; Purvis OW. (2006). Biomonitoring using the lichen Hypogymnia physodes and bark
samples near Zlatna, Romania immediately following closure of a copper ore-processing plant. Environmental Pollution
143: 81-88.
Ryan PA; Lowenthal D; Kumar N. (2005). Improved light extinction reconstruction in interagency monitoring of protected visual
environments. J Air Waste Manag Assoc 55: 1751-1759.
Ryan PH; LeMasters GK. (2007). A Review of Land-use Regression Models for Characterizing Intraurban Air Pollution
Exposure. Inhalation Toxicology 19: 127.
Ryan PH; LeMasters GK; Levin L; Burkle J; Biswas P; Hu S; Grinshpun S; Reponen T. (2008). A land-use regression model for
estimating microenvironmental diesel exposure given multiple addresses from birth through childhood. Science of the
Total Environment.
Ryno M; Rantanen L; Papaioannou E; Konstandopoulos AG; Koskentalo T; Savela K. (2006). Comparison of pressurized fluid
extraction, Soxhlet extraction and sonication for the determination of polycyclic aromatic hydrocarbons in urban air and
diesel exhaust particulate matter. J Environ Monit 8: 488-493.
S Biswas PFMGSHCS. (2005). Performance Evaluation of a Recently Developed Water-Based Condensation Particle Counter.
Aerosol Science and Technology.
Saathoff H; Naumann KH; Schnaiter M; Schock W; Weingartner E; Baltensperger U; Kramer L; Bozoki Z; Poschl U; Niessner
R. (2003). Carbon mass determinations during the AIDA soot aerosol campaign 1999. J Aerosol Sci 34: 1399-1420.
Sabbioni C; Ghedini N; Bonazza A. (2003). Organic anions in damage layers on monuments and buildings. Atmospheric
Environment 37: 1261-1269.
Sabin LD; Lim JH; Stolzenbach KD; Schiff KC. (2005). Contribution of trace metals from atmospheric deposition to stormwater
runoff in a small impervious urban catchment. Water Research 39.
Sabin LD; Lim JH; Stolzenbach KD; Schiff KC. (2006a). Atmospheric dry deposition of trace metals in the coastal region of Los
Angeles, California, USA. Environmental Toxicology and Chemistry 25: 2334-2341.
Sabin LD; Lim JH; Venezia MT; Winer AM; Schiff KC; Stolzenbach KD. (2006b). Dry deposition and resuspension of particle-
associated metals near a freeway in Los Angeles. Atmospheric Environment 40: 7528-7538.
Sadezky A; Muckenhuber H; Grothe H; Niessner R; Poschl U. (2005). Raman microspectroscopy of soot and related
carbonaceous materials: Spectral analysis and structural information. Carbon 43: 1731-1742.
Safar M; Chamiot-Clerc P; Dagher G; Renaud JF. (2001). Pulse pressure, endothelium function, and arterial stiffness in
spontaneously hypertensive rats. Hypertension 38: 1416-1421.
Sagiv SK; Mendola P; Loomis D; Herring AH; Neas LM; Savitz DA; Poole C. (2005). A time-series analysis of air pollution and
preterm birth in Pennsylvania, 1997-2001. Environ Health Perspect 113: 602-606.
Sahlodin AM; Sotudeh-Gharebagh R; Zhu Y. (2007). Modeling of dispersion near roadways based on the vehicle-induced
turbulence concept. Atmospheric Environment 41: 92-102.
Sakagami M; Byron PR; Venitz J; Rypacek F. (2002). Solute disposition in the rat lung in vivo and in vitro: determining regional
absorption kinetics in the presence of mucociliary escalator. J Pharm Sci 91: 594-604.
December 2008
R-60
DRAFT-DO NOT QUOTE OR CITE
-------
Sakamoto N; Hayashi S; Gosselink J; Ishii H; Ishimatsu Y; Mukae H; Hogg JC; van Eeden SF. (2007). Calcium dependent and
independent cytokine synthesis by air pollution particle-exposed human bronchial epithelial cells. Toxicology and applied
pharmacology 225: 134-141.
Salam MT; Millstein J; Li YF; Lurmann FW; Margolis HG; Gilliland FD. (2005). Birth outcomes and prenatal exposure to
ozone, carbon monoxide, and particulate matter: results from the Children's Health Study. Environ Health Perspect 113:
1638-1644.
Saldiva PH; Clarke RW; Coull BA; Stearns RC; Lawrence J; Murthy GG; Diaz E; Koutrakis P; Suh H; Tsuda A; Godleski JJ.
(2002). Lung inflammation induced by concentrated ambient air particles is related to particle composition. American
journal of respiratory and critical care medicine 165: 1610-1617.
Salemaa M; Derome J; Helmisaari HS; Nieminen T; Vanha-Majamaa I. (2004). Element accumulation in boreal bryophytes,
lichens and vascular plants exposed to heavy metal and sulfur deposition in Finland. Science of the Total Environment
324: 141-160.
Salma I; Maenhaut W; Ocskay R; Chi X. (2007). Sampling artefacts, concentration and chemical composition of fine water-
soluble organic carbon and humic-like substances in a continental urban atmospheric environment. Atmospheric
Environment 41:4106-4118.
Salminen K; Karlsson V. (2003). Comparability of low-volume PM10 sampler with β-attenuation monitor in background
air. Atmospheric Environment 37: 3707-3712.
Salonen JT; Salonen R. (1991). Ultrasonographically assessed carotid morphology and the risk of coronary heart disease.
Arterioscler Thromb 11: 1245-1249.
Salvi S; Blomberg A; Rudell B; Kelly F; Sandstrom T; Holgate ST; Frew A. (1999). Acute inflammatory responses in the
airways and peripheral blood after short-term exposure to diesel exhaust in healthy human volunteers. Am J Respir Crit
Care Med 159: 702-709.
Samet JM; Graff D; Berntsen J; Ghio AJ; Huang YC; Devlin RB. (2007). A comparison of studies on the effects of controlled
exposure to fine, coarse and ultrafine ambient particulate matter from a single location. Inhal Toxicol 19 Suppl 1: 29-32.
Samet JM; Zeger SL; Dominici F; Curriero F; Coursac I; Dockery DW; Schwartz J; Zanobetti A. (2000). The National
Morbidity, Mortality, and Air Pollution Study. Part II: Morbidity and mortality from air pollution in the United States.
Res Rep Health Eff Inst 94: 5-70; discussion 71-79.
Samoli E; Analitis A; Touloumi G; Schwartz J; Anderson HR; Sunyer J; Bisanti L; Zmirou D; Vonk JM; Pekkanen J; Goodman
P; Paldy A; Schindler C; Katsouyanni K. (2005). Estimating the exposure-response relationships between particulate
matter and mortality within the APHEA multicity project. Environ Health Perspect 113: 88-95.
Sanderson EG; Farant JP. (2004). Indoor and outdoor polycyclic aromatic hydrocarbons in residences surrounding a Soderberg
aluminum smelter in Canada. EnvironSci Technol 38: 5350-5356.
Santarpia JL; Li RJ; Collins DR. (2004). Direct measurement of the hydration state of ambient aerosol populations. J Geophy s
Res 109: D18209.
Sardar SB; Fine PM; Sioutas C. (2005). Seasonal and spatial variability of the size-resolved chemical composition of participate
matter (PM10) in the Los Angeles Basin. Journal of Geophysical Research D: Atmospheres 110: 1-14.
Sardar SB; Solomon PA; Geller MD; Sioutas C. (2006). Development and evaluation of a high-volume dichotomous sampler for
chemical speciation of coarse and fine particles. J Aerosol Sci 37: 1455-1466.
Sarnat JA; Brown KW; Schwartz J; Coull BA; Koutrakis P. (2005). Ambient gas concentrations and personal particulate matter
exposures: implications for studying the health effects of particles. Epidemiology 16: 385-395.
Sarnat JA; Long CM; Koutrakis P; Coull BA; Schwartz J; Suh HH. (2002). Using Sulfur as a Tracer of Outdoor Fine Particulate
Matter. Environmental science & technology 36: 5305-5314.
Sarnat JA; Marmur A; Klein M; Kim E; Russell AG; Sarnat SE; Mulholland JA; Hopke PK; Tolbert PE. (2008). Fine particle
sources and cardiorespiratory morbidity: an application of chemical mass balance and factor analytical source-
apportionment methods. Environ Health Perspect 116: 459-466.
Sarnat JA; Schwartz J; Catalano PJ; Suh HH. (2001). Gaseous Pollutants in Particulate Matter Epidemiology: Confounders or
Surrogates? Environmental health perspectives 109: 1053-1061.
Sarnat SE; Coull BA; Ruiz PA; Koutrakis P; Suh HH. (2006b). The influences of ambient particle composition and size on
particle infiltration in Los Angeles, CA, residences. J Air Waste ManagAssoc 56: 186-196.
Sarnat SE; Coull BA; Schwartz J; Gold DR; Suh HH. (2006a). Factors affecting the association between ambient concentrations
and personal exposures to particles and gases. EnvironHealth Perspect 114: 649-654.
Sarnat SE; Suh HH; Coull BA; Schwartz J; Stone PH; Gold DR. (2006c). Ambient particulate air pollution and cardiac
arrhythmia in a panel of older adults in Steubenville, Ohio. Occupational and environmental medicine 63: 700-706.
Sato H; Suzuki KT; Sone H; Yamano Y; Kagawa J; Aoki Y. (2003a). DNA-adduct formation in lungs, nasal mucosa, and livers
of rats exposed to urban roadside air in Kawasaki City, Japan. Environmental Research 93: 36-44.
Sato M; Hansen J; Koch D; Lucis A; Ruedy R; Dubovik O; Holben B; Chin M; Novakov T. (2003b). Global atmospheric black
carbon inferred from AAEONET. Proc Natl Acad Sci 100: 6319-6324.
Sauve S. (2001). Speciation of metals in soils. In Allen HE (Ed.), Bioavailability of Metals In Terrestrial Ecosystems: Importance
of Partitioning for Bioavailability to Invertebrates, Microbes and Plants (pp. 158). Pensacola, FL: SETAC Press.
Sawant AA; Cocker DR, 3rd; Miller JW; Taliaferro T; Diaz-Sanchez D; Linn WS; Clark KW; Gong H, Jr. (2008). Generation
and characterization of diesel exhaust in a facility for controlled human exposures. J Air Waste Manag Assoc 58: 829-
837.
December 2008
R-61
DRAFT-DO NOT QUOTE OR CITE
-------
Schafer RB; Paschke A; Vrana B; Mueller R; Liess M. (2008). Performance of the Chemcatcher@u(R) passive sampler when
used to monitor 10 polar and semi-polar pesticides in 16 Central European streams, and comparison with two other
sampling methods. Water Research 42: 2707-2717.
Schanker LS; Mitchell EW; Brown RA. (1986). Species comparison of drug absorption from the lung after aerosol inhalation or
intratracheal injection. Drug Metabolism and Disposition 14: 79-88.
Schatz M; Zeiger RS; Hoffman CP. (1990). Intrauterine growth is related to gestational pulmonary function in pregnant asthmatic
women. Kaiser-Permanente Asthma and Pregnancy Study Group. Chest 98: 389-392.
Schauer C; Niessner R; Poschl U. (2003). Polycyclic aromatic hydrocarbons in urban air particulate matter: decadal and seasonal
trends, chemical degradation, and sampling artifacts. Environ Sci Technol 37: 2861-2868.
Schauer JJ; Kleeman MJ; Cass GR; Simoneit BRT. (1999). Measurement of Emissions from Air Pollution Sources. 2. C~ 1
through C~ 3~ 0 Organic Compounds from Medium Duty Diesel Trucks. ENVIRONMENTAL SCIENCE AND
TECHNOLOGY-WASHINGTON DC- 33: 1578-1587.
Schauer JJ; Kleeman MJ; Cass GR; Simoneit BRT. (2002). Measurement of Emissions from Air Pollution Sources. 5. C~ 1-C~
3~ 2 Organic Compounds from Gasoline-Powered Motor Vehicles. ENVIRONMENTAL SCIENCE AND
TECHNOLOGY-WASHINGTON DC- 36: 1169-1180.
Schauer JJ; Rogge WF; Hildemann LM; Mazurek MA; Cass GR; Simoneit BRT. (1996). Source apportionment of airborne
particulate matter using organic compounds as tracers. Atmospheric Environment 30: 3837-3855.
Schaumann F; Borm PJA; Herbrich A; Knoch J; Pitz M; Schins RPF; Luettig B; Hohlfeld JM; Heinrich J; Krug N. (2004).
Metal-rich ambient particles (Particulate Matter(2.5)) cause airway inflammation in healthy subjects. Am J Respir Crit
Care Med 170: 898-903.
Scheuch G; Gebhart J; Roth C. (1990). Uptake of electrical charges in the human respiratory tract during exposure to air loaded
with negative ions. J Aerosol Sci 21: S439-S442.
Schichtel BA; Gebhart KA; Barna MG; Malm WC; Green MC. (2004). Big Bend Regional Aerosol and Visibility Observational
(BRAVO) Study Results: Air Quality Data and Source Attribution Analyses Results from the National Park Service.
Cooperative Institute for Research in the Atmosphere Colorado State University CIRA, Ft Collins, CO.
Schichtel BA; Gebhart KA; Malm WC; Barna MG; Pitchford ML; Knipping EM; Tombach IH. (2005). Reconciliation and
interpretation of big bend national park particulate sulfur source apportionment: Results from the big bend regional
aerosol and visibility observational study-part I. Journal of the Air & Waste Management Association 55: 1709-1725.
Schichtel BA; Malm WC; Bench G; Fallon S; McDade CE; Chow JC; Watson JG. (2008). Fossil and contemporary fine
particulate carbon fractions at 12 rural and urban sites in the United States. Journal of Geophysical Research-
Atmospheres 113.
Schiff K; Bay S; Stransky C. (2002). Characterization of stormwater toxicants from an urban watershed to freshwater and marine
organisms. Urban Water 4: 215-227.
Schikowski T; Sugiri D; Ranft U; Gehring U; Heinrich J; Wichmann HE; Kramer U. (2005). Long-term air pollution exposure
and living close to busy roads are associated with COPD in women. Respir Res 6: 9921-9926.
Schildcrout JS; Sheppard L; Lumley T; Slaughter JC; Koenig JQ; Shapiro GG. (2006). Ambient air pollution and asthma
exacerbations in children: an eight-city analysis. Am J Epidemiol 164: 505-517.
Schilling JS; Lehman ME. (2002). Bioindication of atmospheric heavy metal deposition in the Southeastern US using the moss
Thuidium delicatulum. Atmospheric Environment 36: 1611-1618.
Schins RP; Lightbody JH; Borm PJ; Shi T; Donaldson K; Stone V. (2004). Inflammatory effects of coarse and fine particulate
matter in relation to chemical and biological constituents. Toxicol Appl Pharmacol 195: 1-11.
Schins RPF; Knaapen AM. (2007). Genotoxicity of poorly soluble particles. Inhalation toxicology 19 Suppl 1: 189-198.
Schnelle-Kreis J; Sklorz M; Peters A; Cyrys J; Zimmermann R. (2005). Analysis of particle-associated semi-volatile aromatic
and aliphatic hydrocarbons in urban particulate matter on a daily basis. Atmospheric Environment 39: 7702-7714.
Schreuder AB; Larson TV; Sheppard L; Claiborn CS. (2006). Ambient woodsmoke and associated respiratory emergency
department visits in Spokane, Washington. International journal of occupational and environmental health 12: 147-153.
Schroeder WH; Munthe J. (1998). Atmospheric mercury—an overview. Atmospheric Environment 32: 809-822.
Schroeter JD; Kimbell JS; Asgharian B. (2006). Analysis of Particle Deposition in the Turbinate and Olfactory Regions Using a
Human Nasal Computational Fluid Dynamics Model. J Aerosol Med 19: 301-313.
Schulz M; Kinne S; Textor C; Guibert S (2004). AeroCom Aerosol Model Intercomparison. Journal.
doi:http://nansen.ipsl.jussieu.fr/AEROCOM/
Schwab JJ; Felton HD; Rattigan OV; Demeijian KL. (2006). New York State Urban and Rural Measurements of Continuous
PM2. 5 Mass by FDMS TEOM and BAM: Evaluations and Comparisons with the FRM. JAWMA 56: 372-383.
Schwab JJ; Hogrefe O; Demerjian KL; Ambs JL. (2004). Laboratory characterization of modified tapered element oscillating
microbalance samplers. J Air Waste Manag Assoc 54: 1254-1263.
Schwartz J. (2003). Airborne Particles and Daily Deaths in 10 US Cities. Revised analyses of time-series studies of air pollution
and health: 211-218.
Schwartz J. (2004a). Air pollution and children's health. Pediatrics 113: 1037-1043.
Schwartz J. (2004b). The effects of particulate air pollution on daily deaths: a multi-city case crossover analysis. Occupational
and environmental medicine 61: 956-961.
Schwartz J. (2004c). Is the Association of Airborne Particles with Daily Deaths Confounded by Gaseous Air Pollutants? An
Approach to Control by Matching. Environmental health perspectives 112: 557-561.
December 2008
R-62
DRAFT-DO NOT QUOTE OR CITE
-------
Schwartz J; Coull B; Laden F; Ryan L. (2008). The effect of dose and timing of dose on the association between airborne
particles and survival. Environ Health Perspect 116: 64-69.
Schwartz J; Litonjua A; Suh H; Verrier M; Zanobetti A; Syring M; Nearing B; Verrier R; Stone P; MacCallum G; Speizer FE;
Gold DR. (2005b). Traffic related pollution and heart rate variability in a panel of elderly subjects. Thorax 60: 455-461.
Schwartz J; Morris R. (1995). Air pollution and hospital admissions for cardiovascular disease in Detroit, Michigan. Am J
Epidemiol 142: 23-35.
Schwartz J; Neas LM. (2000). Fine particles are more strongly associated than coarse particles with acute respiratory health
effects in schoolchildren. Epidemiology 11: 6-10.
Schwartz J; Park SK; O'Neill MS; Vokonas PS; SparrowD; Weiss S; Kelsey K. (2005a). Glutathione-S-Transferase Ml, Obesity,
Statins, and Autonomic Effects of Particles: Gene-by-Drug-by-Environment Interaction. Am J Respir Crit Care Med 172:
1529-1533.
Schwartz J; Sarnat JA; Coull BA; Wilson WE. (2007). Effects of exposure measurement error on particle matter epidemiology: a
simulation using data from a panel study in Baltimore, MD. Journal of Exposure Science and Environmental
Epidemiology 17: S2.
Schwartz J; Zanobetti A; Bateson TF. (2003). Morbidity and mortality among elderly residents in cities with daily PM
measurements. In Revised Analyses of Time-Series Studies of Air Pollution and Health. Special Report, (pp. 25-58).
Boston: Health Effects Institute.
Seagrave J; Dunaway S; McDonald JD; Mauderly JL; Hayden P; Stidley C. (2007). Responses of differentiated primary human
lung epithelial cells to exposure to diesel exhaust at an air-liquid interface. Experimental lung research 33: 27-51.
Seagrave J; Gigliotti A; McDonald JD; Seilkop SK; Whitney KA; Zielinska B; Mauderly JL. (2005a). Composition, toxicity, and
mutagenicity of particulate and semivolatile emissions from heavy-duty compressed natural gas-powered vehicles.
Toxicol Sci 87: 232-241.
Seagrave J; Knall C; McDonald JD; Mauderly JL. (2004). Diesel particulate material binds and concentrates a proinflammatory
cytokine that causes neutrophil migration. Inhalation toxicology 16 Suppl 1: 93-98.
Seagrave J; Mauderly JL; Seilkop SK. (2003). In vitro relative toxicity screening of combined particulate and semivolatile
organic fractions of gasoline and diesel engine emissions. Journal of toxicology and environmental health Part A 66:
1113-1132.
Seagrave J; McDonald JD; Bedrick E; Edgerton ES; Gigliotti AP; Jansen JJ; Ke L; Naeher LP; Seilkop SK; Zheng M; Mauderly
JL. (2006). Lung toxicity of ambient particulate matter from southeastern US sites with different contributing sources:
Relationships between composition and effects. Environmental health perspectives 114: 1387-1393.
Seagrave J; McDonald JD; Reed MD; Seilkop SK; Mauderly JL. (2005b). Responses to subchronic inhalation of low
concentrations of diesel exhaust and hardwood smoke measured in rat bronchoalveolar lavage fluid. Inhalation toxicology
17: 657-670.
Seaman NL. (2000). Meteorological modeling for air-quality assessments. Atmospheric Environment 34: 2231-2259.
Seaton A; MacNee W; Donaldson K; Godden D. (1995). Particulate air pollution and acute health effects. Lancet 345: 176-178.
Seemayer NH; Hornberg C. (1998). Malignant transformation of Syrian hamster kidney cells in vitro by interaction of airborne
particulates and simian virus (SV-) 40. Toxicology Letters 96: 231-238.
Segala C; Poizeau D; Neukirch F; Aubier M; Samson J; Gehanno P. (2004). Air pollution, passive smoking, and respiratory
symptoms in adults. Archives of environmental health 59: 669-676.
Seigneur C; Johnson CD; Latimer DA; Bergstrom RW; Hogo H. (1984). Users manual for the Plume Visibility Model (PLUVUE
II). Final report 23 Feb-29 Aug 83. PB-84-158302, Systems Applications, Inc., San Rafael, CA (USA),
Seigneur C; Lohman K; Vijayaraghavan K; Shia R-L. (2003). Contributions of global and regional sources to mercury deposition
in New York State. Environmental Pollution 123: 365-373.
Seinfeld JH; Pandis SN. (1998). Atmospheric Chemistry and Physics: From Air Pollution to Climate Change. New York, USA.
Seinfeld JH; Pankow JF. (2003). Organic Atmospheric Particulate Material. Annual Review of Physical Chemistry 54: 121-140.
Selevan SG; Borkovec L; Slott VL; Zudova Z; Rubes J; Evenson DP; Perreault SD. (2000). Semen quality and reproductive
health of young Czech men exposed to seasonal air pollution. Environ Health Perspect 108: 887-894.
Semmler M; Seitz J; Erbe F; Mayer P; Heyder J; Oberdorster G; Kreyling WG. (2004). Long-Term Clearance Kinetics of Inhaled
Ultrafme Insoluble Iridium Particles from the Rat Lung, Including Transient Translocation into Secondary Organs.
Inhalation Toxicology 16: 453-459.
Semmler-Behnke M; Takenaka S; Fertsch S; Wenk A; Seitz J; Mayer P; Oberdorster G; Kreyling WG. (2007). Efficient
Elimination of Inhaled Nanoparticles from the Alveolar Region: Evidence for Interstitial Uptake and Subsequent
Reentrainment onto Airways Epithelium. Environmental health perspectives 115: 728.
Serre ML; Christakos G. (1999). Modern geostatistics: computational BME analysis in the light of uncertain physical knowledge-
the Equus Beds study. Stochastic Environmental Research and Risk Assessment (SERRA) 13: 1-26.
Sevastyanova O; Binkova B; Topinka J; Sram RJ; Kalina I; Popov T; Novakova Z; Farmer PB. (2007). In vitro genotoxicity of
PAH mixtures and organic extract from urban air particles part II: human cell lines. Mutat Res 620: 123-134.
Shah AP; Pietropaoli AP; Frasier LM; Speers DM; Chalupa DC; Delehanty JM; Huang LS; Utell MJ; Frampton MW. (2008).
Effect of inhaled carbon ultrafme particles on reactive hyperemia in healthy human subjects. Environ Health Perspect
116: 375-380.
Shakya K; Chettri MK; Sawidis T. (2008). Impact of heavy metals (copper, zinc, and lead) on the chlorophyll content of some
mosses. Archives of Environmental Contamination and Toxicology 54: 412-421.
December 2008
R-63
DRAFT-DO NOT QUOTE OR CITE
-------
Shan XQ; Wang ZW; Wang WS; Zhang SZ; Wen B. (2003). Labile rhizosphere soil solution fraction for prediction of
bioavailability of heavy metals and rare earth elements to plants. Annals of Bioanalytical Chemistry 375: 400-407.
Shantz NC; Leaitch WR; Caffrey PF. (2003). Effect of organics of low solubility on the growth rate of cloud droplets. Journal of
Geophysical Research 108: 4168.
Sharifi MR; Gibson AC; Rundel PW. (1999). Pheno logical and physiological responses of heavily dusted creosote bush (Larrea
tridentata) to summer irrigation in the Mojave Desert. Flora 14: 369-378.
Sharma AK; Jensen KA; Rank J; White PA; Lundstedt S; Gagne R; Jacobsen NR; Kristiansen J; Vogel U; Wallin H. (2007).
Genotoxicity, inflammation and physico-chemical properties of fine particle samples from an incineration energy plant
and urban air. Mutat Res 633: 95-111.
Sharma M; Kumar VN; Katiyar SK; Sharma R; Shukla BP; Sengupta B. (2004). Effects of particulate air pollution on the
respiratory health of subjects who live in three areas in Kanpur, India. Archives of environmental health 59: 348-358.
Shaw LJ; Raggi P; Schisterman E; Berman DS; Callister TQ. (2003). Prognostic value of cardiac risk factors and coronary artery
calcium screening for all-cause mortality. Radiology 228: 826-833.
Sheesley RJ. (2007). Daily Variation in Particle-Phase Source Tracers in an Urban Atmosphere. Aerosol Science and Technology
41: 981-993.
Sheesley RJ; Schauer JJ; Elemming JD; Barman MA; Geis SW; Tortorelli JJ. (2004). Toxicity of ambient atmospheric particulate
matter from the Lake Michigan (USA) airshed to aquatic organisms. Environmental Toxicology and Chemistry 23: 133-
140.
Shemesh J; Tenenbaum A; Fisman EZ; Apter S; Rath S; Rozenman J; Itzchak Y; Motro M. (1996). Absence of coronary
calcification on double-helical CT scans: predictor of angiographically normal coronary arteries in elderly women?
Radiology 199: 665-668.
Shen S; Jaques PA; Zhu Y; Geller MD; Sioutas C. (2002). Evaluation of the SMPS-APS system as a continuous monitor for
measuring PM2. 5, PM10 and coarse (PM2. 5-10) concentrations. Atmospheric Environment 36: 3939-3950.
Sheppard L; Slaughter JC; Schildcrout J; Liu LJS; Lumley T. (2005). Exposure and measurement contributions to estimates of
acute air pollution effects. Journal of exposure analysis and environmental epidemiology 15: 366-376.
Sheya SA; Glowacki C; Chang MC; Chow JC; Watson JG. (2008). Hot filter/impinger and dilution sampling for fine particulate
matter characterization from ferrous metal casting processes. J Air Waste Manag Assoc 58: 553-561.
Shi JP; Harrison RM; D. E. (2001). Comparison of ambient particle surface area measurement by epiphaniometer and
SMPS/APS. Atmospheric Environment 35: 6193-6200.
Shi T; Knaapen AM; Begerow J; Birmili W; Borm PJA; Schins RPF. (2003). Temporal variation of hydroxyl radical generation
and 8-hydroxy-2'-deoxyguanosine formation by coarse and fine particulate matter. Occup Environ Med 60: 315-321.
Shinkura R; Fijuyama C; Akiba S. (1999). Relationship between ambient sulfur dioxide levels and neonatal mortality near the
Mt. Sakurajima volcano in Japan. J Epidemiol 9: 344-349.
Shinn JH. (1978). A critical survey of measurements of foliar deposition of airborne sulfates and nitrates. Paper presented at the
71st annual meeting of the Air Pollution Control Association; June; Houston, TX, Pittsburgh, PA.
Shrivastava MK; Subramanian R; Rogge WF; Robinson AL. (2007). Sources of organic aerosol: Positive matrix factorization of
molecular marker data and comparison of results from different source apportionment models. Atmospheric Environment
41: 9353-9369.
Shulz KG; Zondervan I; Gerringa LJA; Timmermans KR; Veldhuls MJW; Riebesell U. (2004). Effect of trace metal availability
on coccolithophorid calcification. Nature 430: 673-676.
Silkoff PE; Zhang L; Dutton S; Langmack EL; Vedal S; Murphy J; Make B. (2005). Winter air pollution and disease parameters
in advanced chronic obstructive pulmonary disease panels residing in Denver, Colorado. J Allergy Clin Immunol 115:
337-344.
Sillanpaa M; Frey A; Hillamo R; Pennanen AS; Salonen RO. (2005). Organic, elemental and inorganic carbon in particulate
matter of six urban environments in Europe. Atmos Chem Phys 5: 2869-2879.
Sillanpaa M; Hillamo R; Saarikoski S; Frey A; Pennanen A; Makkonen U; Spolnik Z; Van Grieken R; Branis M; Brunekreef B;
Chalbot M-C; Kuhlbusch T; Sunyer J; Kerminen V-M; Kulmala M; Salonen RO. (2006). Chemical composition and mass
closure of particulate matter at six urban sites in Europe. Atmospheric Environment 40: 212-223.
Silva PJ; Erupe ME; Price D; Elias J; Malloy QG; Li Q; Warren B; Cocker DR, 3rd. (2008). Trimethylamine as precursor to
secondary organic aerosol formation via nitrate radical reaction in the atmosphere. Environ Sci Technol 42: 4689-4696.
Simons E; Curtin-Brosnan J; Buckley T; Breysse P; Eggleston PA. (2007). Indoor environmental differences between inner city
and suburban homes of children with asthma. J Urban Health 84: 577-590.
Simpson R; Williams G; Petroeschevsky A; Best T; Morgan G; Denison L; Hinwood A; Neville GCEH. (2005). The short-term
effects of air pollution on hospital admissions in four Australian cities. Aust N Z J Public Health 29:213-221.
Sinclair AH; Tolsma D. (2004). Associations and lags between air pollution and acute respiratory visits in an ambulatory care
setting: 25-month results from the aerosol research and inhalation epidemiological study. J Air Waste Manag Assoc 54:
1212-1218.
Singh M; Misra C; Sioutas C. (2003). Field evaluation of a personal cascade impactor sampler (PCIS). Atmospheric Environment
37: 4781-4793.
Sioutas C; Delfino RJ; Singh M. (2005). Exposure Assessment for Atmospheric Ultrafme Particles (UFPs) and Implications in
Epidemiologic Research. Environmental health perspectives 113: 947.
December 2008
R-64
DRAFT-DO NOT QUOTE OR CITE
-------
Sioutas C; Kim S; Chang M. (1999). Development and evaluation of a prototype ultrafme particle concentrator. J Aerosol Sci 30:
1001-1017.
Sioutas C; Koutrakis P; Ferguson ST; Burton RM. (1995). Development and Evaluation of a Prototype Ambient Particle
Concentrator for Inhalation Exposure Studies. Inhalation Toxicology 7: 633-644.
Sioutas C; Koutrakis P; Godleski JJ; Ferguson ST; Kim CS; Burton RM. (1997). Fine particle concentrators for inhalation
exposures - Effect of particle size and composition. J Aerosol Sci 28: 1057-1071.
Sirivelu MP; MohanKumar SMJ; Wagner JG; Harkema JR; MohanKumar PS. (2006). Activation of the stress axis and
neurochemical alterations in specific brain areas by concentrated ambient particle exposure with concomitant allergic
airway disease. Environmental health perspectives 114: 870-874.
Sisler JF; Malm WC; Gebhart KA; Colo Cooperative Institute for Research in the A. (1996). Spatial and Seasonal Patterns and
Long Term Variability of the Composition of the Haze in the United States: An Analysis of Data from the IMPROVE
Network: Cooperative Institute for Research in the Atmosphere, Colorado State University.
Slama R; Darrow L; Parker J; Woodruff TJ; Strickland M; Nieuwenhuijsen M; Glinianaia S; Hoggatt KJ; Kannan S; Hurley F;
Kalinka J; Sram R; Brauer M; Wilhelm M; Heinrich J; Ritz B. (2008). Meeting report: atmospheric pollution and human
reproduction. Environ Health Perspect 116: 791-798.
Slama R; Morgenstern V; Cyrys J; Zutavern A; Herbarth O; Wichmann HE; Heinrich J. (2007). Traffic-related atmospheric
pollutants levels during pregnancy and offspring's term birth weight: a study relying on a land-use regression exposure
model. Environ Health Perspect 115: 1283-1292.
Slaughter JC; Kim E; Sheppard L; Sullivan JH; Larson TV; Claiborn C. (2005). Association between particulate matter and
emergency room visits, hospital admissions and mortality in Spokane, Washington. Journal of exposure analysis and
environmental epidemiology 15: 153-159.
Slaughter JC; Lumley T; Sheppard L; Koenig JQ; Shapiro GG. (2003). Effects of ambient air pollution on symptom severity and
medication use in children with asthma. Ann Allergy Asthma Immunol 91: 346-353.
Sterdal LH, Walker, S.-E. and Solberg, S. (2003). The urban air dispersion model EPISODE applied in AirQUIS2003. Technical
description. Norwegian Institute for Air Research, Kjeller. RT 12/2003.
Slowik JG. (2007). An Inter-Comparison of Instruments Measuring Black Carbon Content of Soot Particles. Aerosol Science and
Technology 41: 295-314.
Smejkalova M; Mikanova O; Boruvka L. (2003). Effects of heavy metal concentrations on biological activity of soil
microorganisms. Plant Soil Environ 49: 321-326.
Smilde TJ; Wollersheim H; Van Langen H; Stalenhoef AF. (1997). Reproducibility of ultrasonographic measurements of
different carotid and femoral artery segments in healthy subjects and in patients with increased intima-media thickness.
Clin Sci (Lond) 93: 317-324.
Smith JN; Moore KF; McMurry PH; Eisele FL. (2004). Atmospheric Measurements of Sub-20 nm Diameter Particle Chemical
Composition by Thermal Desorption Chemical Ionization Mass Spectrometry. Aerosol Science and Technology 38: 100-
110.
Smith KR; Kim S; Recendez JJ; Teague SV; Menache MG; Grubbs DE; Sioutas C; Pinkerton KE. (2003). Airborne particles of
the California central valley alter the lungs of healthy adult rats. Environmental health perspectives 111: 902-908.
Smith KR; Veranth JM; Kodavanti UP; Aust AE; Pinkerton KE. (2006). Acute pulmonary and systemic effects of inhaled coal
fly ash in rats: comparison to ambient environmental particles. Toxicological sciences 93: 390-399.
Smith RL; Spitzner D; Kim Y; Fuentes M. (2000). Threshold dependence of mortality effects for fine and coarse particles in
Phoenix, Arizona. J Air Waste Manag Assoc 50: 1367-1379.
Smith W. (1990). Forest nutrient cycling: toxic ions. In Air Pollution and Forests: Interactions between Air Contaminants and
Forest Ecosystems (pp. 225-268). New York, NY: Springer-Verlang.
Smith WH. (1991). Air pollution and forest damage. Chemical & Engineering News 69: 30-43.
Smith WH; Staskawicz BJ. (1977). Removal of atmospheric particles by leaves and twigs of urban trees: some preliminary
observations and assessment of research needs. Environmental Management 1: 317-330.
Smolders E; Degryse F. (2002). Fate and effect of zinc from tire debris in soil. Environmental Science and Technology 36: 3706-
3710.
Snipes MB. (1996). Current Information on Lung Overload in Nonrodent Mammals: Contrast with Rats. Inhalation Toxicology
8: 91-109.
Snipes MB; Harkema JR; Hotchkiss JA; Bice DE. (1997). Neutrophil Involvement in the Retention and Clearance of Dust
Intratracheally Instilled into the LUNGS of F344/N Rats. Experimental lung research 23: 65-84.
So ESP; Chan ATY; Wong AYT. (2005). Large-eddy simulations of wind flow and pollutant dispersion in a street canyon.
Atmospheric Environment 39: 3573-3582.
Soares SR; Bueno-Guimaraes HM; Ferreira CM; Rivero DH; De Castro I; Garcia ML; Saldiva PH. (2003). Urban air pollution
induces micronuclei in peripheral erythrocytes of mice in vivo. Environ Res 92: 191-196.
Soberanes S; Panduri V; Mutlu GM; Ghio A; Bundinger GRS; Kamp DW. (2006). p53 mediates particulate matter-induced
alveolar epithelial cell mitochondria-regulated apoptosis. Am J Respir Crit Care Med 174: 1229-1238.
Soderholm SC. (1985). Size-selective sampling criterial for inspirable mass fraction. Particle size-selective sampling in the
workplace: 27-32.
Sokol RZ; Kraft P; Fowler IM; Mamet R; Kim E; Berhane KT. (2006). Exposure to environmental ozone alters semen quality.
Environ Health Perspect 114: 360-365.
December 2008
R-65
DRAFT-DO NOT QUOTE OR CITE
-------
Solomon P; Baumann K; Edgerton E; Tanner R; Eatough D; Modey W; Marin H; Savoie D; Natarajan S; Meyer MB. (2003).
Comparison of integrated samplers for mass and composition during the 1999 Atlanta supersites project. J Geophys Res
108: 8423.
Solomon PA; Hopke PK. (2008). A Special Issue of JA&WMA Supporting Key Scientific and Policy-and Health-Relevant
Findings from EPA's Particulate Matter Supersites Program and Related Studies: An Integration and Synthesis of Results.
JOURNAL-AIR AND WASTE MANAGEMENT ASSOCIATION 58: 137.
Somers CM; McCarry BE; Malek F; Quinn JS. (2004). Reduction of particulate air pollution lowers the risk of heritable
mutations in mice. Science 304: 1008-1010.
Somers CM; Yauk CL; White PA; Parfett CL; Quinn JS. (2002). Air pollution induces heritable DNA mutations. Proc Natl Acad
Sci U S A 99: 15904-15907.
Song CL; Zhou YC; Huang RJ; Wang YQ; Huang QF; Lu G; Liu KM. (2007). Influence of ethanol-diesel blended fuels on diesel
exhaust emissions and mutagenic and genotoxic activities of particulate extracts. J Hazard Mater 149: 355-363.
Song H-M; Jang A-S; Ahn M-H; Takizawa H; Lee S-H; Kwon J-H; Lee Y-M; Rhim T; Park C-S. (2008). Yml and Ym2
expression in a mouse model exposed to diesel exhaust particles. Environmental toxicology 23: 110-116.
Sorensen M; Daneshvar B; Hansen M; Dragsted LO; Hertel O; Knudsen L; Loft S. (2003). Personal PM2.5 exposure and markers
of oxidative stress in blood. Environ Health Perspect 111: 161-166.
Sorensen M; Schins RP; Hertel O; Loft S. (2005). Transition metals in personal samples of PM2.5 and oxidative stress in human
volunteers. Cancer Epidemiol Biomarkers Prev 14: 1340-1343.
S0renson M; Loft S; Andersen HV; Raaschou-Nielsen O; Skovgaard LT; Knudsen LE; Nielsen IV; Hertel O. (2005). Personal
exposure to PM2.5, black smoke and N02 in Copenhagen: relationship to bedroom and outdoor concentrations covering
seasonal variation. J ExpoAnalEnvironEpidemiol 15: 413-422.
Sousa SR; Moradas-Ferreira P; Saramago B; Melo LV; Barbosa MA. (2004). Human Serum Albumin Adsorption on TiO~ 2
from Single Protein Solutions and from Plasma. Langmuir 20: 9745-9754.
Sram RJ; Binkova B; Dejmek J; Bobak M. (2005). Ambient air pollution and pregnancy outcomes: a review of the literature.
Environ Health Perspect 113: 375-382.
Stanier C; Khlystov A; Chan W; Mandiro M; Pandis S. (2004). A Method for the In Situ Measurement of Fine Aerosol Water
Content of Ambient Aerosols: The Dry-Ambient Aerosol Size Spectrometer (DAASS). Aerosol Science and Technology
38: 215-228.
Stanier CO; Pathak RK; Pandis SN. (2007). Measurements of the volatility of aerosols from alpha-pinene ozonolysis. Environ Sci
Technol 41: 2756-2763.
Stanojevic S; Wade A; Stocks J; Hankinson J; Coates AL; Pan H; Rosenthal M; Corey M; Lebecque P; Cole TJ. (2008).
Reference ranges for spirometry across all ages: a new approach. American journal of respiratory and critical care
medicine 177: 253-260.
Steerenberg PA; van Amelsvoort L; Lovik M; Hetland RB; Alberg T; Halatek T; Bloemen HJT; Rydzynski K; Swaen G;
Schwarze P; Dybing E; Cassee FR. (2006). Relation between sources of particulate air pollution and biological effect
parameters in samples from four European cities: an exploratory study. Inhalation toxicology 18: 333-346.
Steerenberg PA; Withagen CET; van Dalen WJ; Dormans JAMA; Cassee FR; Heisterkamp SH; van Loveren H. (2004b).
Adjuvant activity of ambient particulate matter of different sites, sizes, and seasons in a respiratory allergy mouse model.
Toxicology and applied pharmacology 200: 186-200.
Steerenberg PA; Withagen CET; van Dalen WJ; Dormans JAMA; Heisterkamp SH; van Loveren H; Cassee FR. (2005). Dose
dependency of adjuvant activity of particulate matter from five European sites in three seasons in an ovalbumin-mouse
model. Inhalation toxicology 17: 133-145.
Steerenberg PA; Withagen CET; van Dalen WJ; Dormans JAMA; van Loveren H. (2004a). Adjuvant activity of ambient
particulate matter in macrophage activity-suppressed, N-acetylcysteine-treated, iNOS- and IL-4-deficient mice. Inhalation
toxicology 16: 835-843.
Steinnes E; Hvatum O0; B0lviken B; Varskog P. (2005). Atmospheric supply of trace elements studied by peat samples from
ombrotrophic bogs. J Environ Qual 34: 192-197.
Steinvil A; Kordova-Biezuner L; Shapira I; Berliner S; Rogowski O. (2008). Short-term exposure to air pollution and
inflammation-sensitive biomarkers. Environ Res 106: 51-61.
Stenfors N; Nordenhall C; Salvi SS; Mudway I; Soderberg M; Blomberg A; Helleday R; Levin JO; Holgate ST; Kelly FJ; Frew
AJ; Sandstrom T. (2004). Different airway inflammatory responses in asthmatic and healthy humans exposed to diesel.
Eur Respir J 23: 82-86.
Stern GA; Halsall CJ; Barrie LA; Muir DCG; Fellin P; Rosenberg B. (1997). Polychlorinated biphenyls in arctic air. 1. Temporal
and spatial trends: 1992-1994. Environmental Science and Technology 31: 3619-3628.
Stevens T; Krantz QT; Linak WP; Hester S; Gilmour MI. (2008). Increased Transcription of Immune and Metabolic Pathways in
Naive and Allergic Mice Exposed to Diesel Exhaust. Toxicol Sci.
Stewart AR; Luoma SN; Schlekat CE; Doblin MA; Hieb KA. (2004). Food web pathway determines how selenium affects
aquatic ecosystems. Environmental Science and Technology 38: 4519-4526.
Stieb DM; Evans GJ; Sabaliauskas K; Chen LI; Campbell ME; Wheeler AJ; Brook JR; Guay M. (2008). A scripted activity study
of the impact of protective advice on personal exposure to ultra-fine and fine particulate matter and volatile organic
compounds. Journal of Exposure Science and Environmental Epidemiology 18: 495-502.
Stier P, et al. (2005). The aerosol-climate model ECHAM5-HAM. Atmospheric Chemistry and Physics 5: 1125-1156.
December 2008
R-66
DRAFT-DO NOT QUOTE OR CITE
-------
Stier P; Seinfeld JH; Kinne S; Boucher O. (2007). Aerosol absorption and radiative forcing. Atmos Chem Phys 7: 5237-5261.
Stacker R; Keaney JF, Jr. (2004). Role of oxidative modifications in atherosclerosis. Physiol Rev 84: 1381-1478.
Stohl A; Andrews E; Burkhart JF; Forster C; Herber A; Hoch SW; Kowal D; Lunder C; Mefford F; Ogren JA; Sharma S;
Spichtinger N; Stebel K; Stone R; Strom J; T0rseth K; Wehrli C; Yttri KE. (2006). Pan-Arctic enhancements of light
absorbing aerosol concentrations due to North American boreal forest fires during summer 2004. Journal of Geophysical
Research 111, D22214, doi:10.1029/2006JD007216.
Stohl A; Berg F; Burkhart JF; Fjasraa AM; Forster C; Herber A; Hov 0; Lunder C; McMillan WW; Oltmans S; Shiobara M;
Dimpson D; Solberg S; Stebel K; Strom J; T0rseth K; Treffeisen R; Virkkunen K; Yttri KE. (2007). Arctic smoke -
record high air pollution levels in the European Arctic due to agricultural fires in Eastern Europe in spring 2006. Atmos
Chem Phys 7: 511-534.
Stolzel M; Breitner S; Cyrys J; Pitz M; Wolke G; Kreyling W; Fleinrich J; Wichmann FIE; Peters A. (2007). Daily mortality and
particulate matter in different size classes in Erfurt, Germany. Journal of Exposure Science and Environmental
Epidemiology 17: 458-467.
Stolzel M; Peters A; Wichmann FIE. (2003). Daily mortality and fine and ultrafme particles in Erfurt, Germany. Revised analyses
of time-series studies of air pollution and health Special report Boston, Massachusetts, Flealth Effects Institute: 231-240.
Stolzenburg MR; Dutcher DD; Kirby BW; Flering SV. (2003). Automated Measurement of the Size and Concentration of
Airborne Particulate Nitrate. Aerosol Science and Fechnology 37: 537-546.
Stolzenburg MR; Flering SV. (2000). Method for the automated measurement of fine particle nitrate in the atmosphere.
Environmental Science and Fechnology 34: 907-914.
Strachan WMJ; Burniston DA; Williamson M; Bohdanowicz H. (2001). Spatial difference in persistent organochlorine pollutant
concentrations between the Bering and Chukchi Seas. Marine Pollution Bulletin 43: 132-142.
Stracquadanio M; Bergamini D; Massaroli E; Frombini C. (2005). Field evaluation of a passive sampler of polycyclic aromatic
hydrocarbons (PAFls) in an urban atmosphere (Bologna, Italy). J Environ Monit 7: 910-915.
Strand M; Hopke PK; Zhao W; Vedal S; Gelfand E; Rabinovitch N. (2007). A study of health effect estimates using competing
methods to model personal exposures to ambient PM2.5. J ExpoSci EnvironEpidemiol 17: 549-558.
Strand M; Vedal S; Rodes C; Dutton SJ; Gelfand EW; Rabinovitch N. (2006). Estimating effects of ambient PM2.5 exposure on
health using PM2.5 component measurements and regression calibration (vol 16, pg 30, 2006). Journal of Exposure
Science and Environmental Epidemiology 16: 471-471.
Strandberg B; Axelsen JA; Pedersen MB; Jensen J; Attrill MJ. (2006). Effect of a copper gradient on plant community structure.
Environmental Foxicology and Chemistry 25: 743-753.
Strandell M; Zakrisson S; Alsberg F; Westerholm R; WinquistL; Rannug U. (1994). Chemical analysis and biological testing of
a polar fraction of ambient air, diesel engine, and gasoline engine particulate extracts. Environmental health perspectives
102: 85—92.
Streets DG; Zhang Q; Wang L; He K; Hao J; Wu Y; Fang Y; Carmichael GR. (2006). Revisiting China's CO emissions after the
Fransport and Chemical Evolution over the Pacific (FRACE-P) mission: Synthesis of inventories, atmospheric modeling,
and observations. J Geophys Res 111.
Strom KA; Garg BD; Johnson JT; DArcy JB; Smiler KL. (1990). Inhaled particle retention in rats receiving low exposures of
diesel exhaust. Journal of Foxicology and Environmental Health 29: 377-398.
Su Y; Sipin MF; Spencer MT; Qin X; Moffet RC; Shields FG; Prather KA; Venkatachari P; Jeong C-H; Kim E; Hopke PK;
Gelein RM; Utell MJ; OberdAT[rster Gn; Berntsen J; Devlin RB; Chen FC. (2006). Real-Fime Characterization of the
Composition of Individual Particles Emitted From Ultrafine Particle Concentrators. Aerosol Science and Fechnology 40:
437-455.
Subramanian R; Khlystov A; Robinson A. (2006). Effect of Peak Inert-Mode Femperature on Elemental Carbon Measured Using
Fhermal-Optical Analysis. Aerosol Science and Fechnology 40: 763-780.
Subramanian R; Khlystov AY; Cabada JC; Robinson AF. (2004). Positive and Negative Artifacts in Particulate Organic Carbon
Measurements with Denuded and Undenuded Sampler Configurations. Aerosol Science and Fechnology 38: 27-48.
Suga K; Yuan Y; Ogasawara N; Fsukuda F; Matsunaga N. (2003). Altered Clearance of Gadolinium
Diethylenetriaminepentaacetic Acid Aerosol from Bleomycin-injured Dog Fungs: Initial Observations. American journal
of respiratory and critical care medicine 167: 1704.
Sugamata M; Ihara F; Fakeda K. (2006). Maternal Exposure to Diesel Exhaust Feads to Pathological Similarity to Autism in
Newborns. Journal of health science 52: 486-488.
Sugiri D; Ranft U; Schikowski F; Kramer U. (2006). Fhe Influence of Farge-Scale Airborne Particle Decline and Fraffic-Related
Exposure on Children's Fung Function. Environmental health perspectives 114: 282.
Suglia SF; Gryparis A; Wright RO; Schwartz J; Wright RJ. (2008). Association of black carbon with cognition among children in
a prospective birth cohort study. Am J Epidemiol 167: 280-286.
Sullivan AP; Peltier RE; Brock CA; de Gouw JA; Holloway JS; Warneke C; Wollny AG; Weber RJ. (2006). Airborne
measurements of carbonaceous aerosol soluble in water over northeastern United States: Method development and an
investigation into water-soluble organic carbon sources. J Geophys Res 111: 1-14.
Sullivan AP; Weber RJ. (2006). Chemical characterization of the ambient organic aerosol soluble in water: 1. Isolation of
hydrophobic and hydrophilic fractions with a XAD-8 resin. J Geophys Res 111.
Sullivan AP; Weber RJ; Clements AF; Furner JR; Bae MS; Schauer JJ. (2004). A method for on-line measurement of water-
soluble organic carbon in ambient aerosol particles: Results from an urban site. Geophys Res Fett 31.
December 2008
R-67
DRAFT-DO NOT QUOTE OR CITE
-------
Sullivan J; Ishikawa N; Sheppard L; Siscovick D; Checkoway H; Kaufman J. (2003). Exposure to Ambient Fine Particulate
Matter and Primary Cardiac Arrest among Persons With and Without Clinically Recognized Heart Disease. Am J
Epidemiol 157: 501-509.
Sullivan J; Sheppard L; Schreuder A; Ishikawa N; Siscovick D; Kaufman J. (2005a). Relation between short-term fme-particulate
matter exposure and onset of myocardial infarction. Epidemiology 16: 41-48.
Sullivan JH; Hubbard R; Liu SL; Shepherd K; Trenga CA; Koenig JQ; Chandler WL; Kaufman ID. (2007). A community study
of the effect of particulate matter on blood measures of inflammation and thrombosis in an elderly population. Environ
Health 6: 3.
Sullivan JH; Schreuder AB; Trenga CA; Liu SLJ; Larson TV; Koenig JQ; Kaufman JD. (2005b). Association between short term
exposure to fine particulate matter and heart rate variability in older subjects with and without heart disease. Thorax 60:
462-466.
Sun Q; Wang A; Jin X; Natanzon A; Duquaine D; Brook RD; Aguinaldo J-GS; Fayad ZA; Fuster V; Lippmann M; Chen Lung C;
Rajagopalan S. (2005). Long-term air pollution exposure and acceleration of atherosclerosis and vascular inflammation in
an animal model. JAMA 294: 3003-3010.
Sun Q; Yue P; Kirk RI; Wang A; Moatti D; Jin X; Lu B; Schecter AD; Lippmann M; Gordon T; Chen LC; Rajagopalan S.
(2008). Ambient air particulate matter exposure and tissue factor expression in atherosclerosis. Inhal Toxicol 20: 127-
137.
Sureshkumar V; Paul B; Uthirappan M; Pandey R; Sahu AP; Lai K; Prasad AK; Srivastava S; Saxena A; Mathur N; Gupta YK.
(2005). Proinflammatory and anti-inflammatory cytokine balance in gasoline exhaust induced pulmonary injury in mice.
Inhal Toxicol 17: 161-168.
Suwa T; Hogg JC; Quinlan KB; Ohgami A; Vincent R; van Eeden SF. (2002). Particulate air pollution induces progression of
atherosclerosis. J Am Coll Cardiol 39: 935-942.
Svartengren M; Falk R; Philipson K. (2005). Long-term clearance from small airways decreases with age (Vol. 26, pp. 609-615):
Eur Respiratory Soc.
Swartz E; Stockburger L; Gundel LA. (2003). Recovery of Semivolatile Organic Compounds during Sample Preparation:
Implications for Characterization of Airborne Particulate Matter. ENVIRONMENTAL SCIENCE AND
TECHNOLOGY-WASHINGTON DC- 37: 597-605.
Swietlik D; Faust M. (1984). Foliar nutrition of fruit crops. Horticulture Review 6: 287-355.
Symons JM; Wang L; Guallar E; Howell E; Dominici F; Schwab M; Ange BA; Samet J; Ondov J; Harrison D; Geyh A. (2006).
A Case-Crossover Study of Fine Particulate Matter Air Pollution and Onset of Congestive Heart Failure Symptom
Exacerbation Leading to Hospitalization. Am J Epidemiol.
Szabo L; Fodor L. (2006). Uptake of microelements by crops grown on heavy metal-amended soil. Communications in Soil
Science and Plant Analysis 37: 2679-2689.
Tabachnik E; Muller N; Toye B; Levison H. (1981). Measurement of ventilation in children using the respiratory inductive
plethysmograph. JPediatr 99: 895-899.
Taha G. (2007). Black Carbon Measurement using Laser Integrating Plate Method. Aerosol Science and Technology 41: 266-
276.
Takenaka S; Dornhofer-Takenaka H; Muhle H. (1986). Alveolar distribution of fly ash and of titanium dioxide after long-term
inhalation by Wistar rats. J Aerosol Sci 17: 361-364.
Takenaka S; Karg E; Kreyling W; Lentner B; Moller W; Behnke-Semmler M; Jennen L; Walch A; Michalke B; Schramel P.
(2006). Distribution Pattern of Inhaled Ultrafine Gold Particles in the Rat Lung. Inhalation Toxicology 18: 733-740.
Takizawa H; Abe S; Okazaki H; Kohyama T; Sugawara I; Saito Y; Ohtoshi T; Kawasaki S; Desaki M; Nakahara K; Yamamoto
K; Matsushima K; Tanaka M; Sagai M; Kudoh S. (2003). Diesel exhaust particles upregulate eotaxin gene expression in
human bronchial epithelial cells via nuclear factor-kappa B-dependent pathway. American journal of physiology Lung
cellular and molecular physiology 284: L1055-1062.
Tal TL; Graves LM; Silbajoris R; Bromberg PA; Wu W; Samet JM. (2006). Inhibition of protein tyrosine phosphatase activity
mediates epidermal growth factor receptor signaling in human airway epithelial cells exposed to Zn2+. Toxicology and
applied pharmacology 214: 16-23.
Tamaoki J; Isono K; Takeyama K; Tagaya E; Nakata J; Nagai A. (2004). Ultrafine carbon black particles stimulate proliferation
of human airway epithelium viaEGF receptor-mediated signaling pathway. American journal of physiology Lung
cellular and molecular physiology 287: LI 127-1133.
Tan WC; Qiu DW; Liam BL; Ng TP; Lee SH; van Eeden SF; D'Yachkova Y; Hogg JC. (2000). The human bone marrow
response to acute air pollution caused by forest fires. Am J Respir Crit Care Med 161: 1213-1217.
Tang IN. (1996). Chemical and size effects of hygroscopic aerosols on light scattering coefficients. J Geophys Res 101: 245-219.
Tankersley CG; Bierman A; Rabold R. (2007). Variation in heart rate regulation and the effects of particle exposure in inbred
mice. Inhalation Toxicology 19: 621-629.
Tankersley CG; Campen M; Bierman A; Flanders SE; Broman KW; Rabold R. (2004). Particle effects on heart-rate regulation in
senescent mice. Inhalation Toxicology 16: 381-390.
Tankersley CG; Champion HC; Takimoto E; Gabrielson K; Bedja D; Misra V; El-Haddad H; Rabold R; Mitzner W. (2008).
Exposure to inhaled particulate matter impairs cardiac function in senescent mice. Am J Physiol Regul Integr Comp
Physiol 295: R252-263.
December 2008
R-68
DRAFT-DO NOT QUOTE OR CITE
-------
Tankersley CG; Irizarry R; Flanders SE; Rabold R; Frank R. (2003). Unstable heart rate and temperature regulation predict
mortality in AKR/J mice. Am J Physiol Regul Integr Comp Physiol 284: R742-750.
Tao F; Gonzalez-Flecha B; Kobzik L. (2003). Reactive oxygen species in pulmonary inflammation by ambient particulates. Free
Radical Biology and Medicine 35: 327-340.
Tao F; Kobzik L. (2002). Lung macrophage-epithelial cell interactions amplify particle-mediated cytokine release. Am J Respir
Cell Mol Biol 26: 499-505.
Tao S; Jiao X; Chen S; Xu F; Li Y; Liu F. (2006). Uptake of vapor and particulate polycyclic aromatic hydrocarbons by cabbage.
Environmental Pollution 140: 13-15.
Tasdemir Y; Kural C; Cindoruk SS; Vardar N. (2006). Assessment of trace element concentrations and their estimated dry
deposition fluxes in an urban atmosphere. Atmospheric Research 81: 17-35.
Tasdemir Y; Odabasi M; Vardar N; Sofuoglu A; Murphy TJ; Holsen TM. (2004). Dry deposition fluxes and velocities of
polychlorinated biphenyls (PCBs) associated with particles. Atmospheric Environment 38: 2447-2456.
Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. (1996).
Heart Rate Variability Standards of Measurement, Physiological Interpretation, and Clinical Use. Circulation 93: 1043-
1065.
Taylor Jr. GE; Hanson PJ; Baldocchi DD. (1988). Pollutant deposition to individual leaves and plant canopies: sites of regulation
and relationship to injury. In Heck WW, Taylor OC, Tingey DT (Eds.), Assessment of crop loss from air pollutants (pp.
227-257). New York, NY: Elsevier Applied Science.
Taylor PE; Flagan RC; Valenta R; Glovsky MM. (2002). Release of allergens as respirable aerosols: A link between grass pollen
and asthma. J Allergy Clin Immunol 109: 51-56.
Telmer K; Bonham-Carter GF; Kliza DA; Hall GEM. (2004). The atmospheric transport and deposition of smelter emissions:
Evidence from the multi-element geochemistry of snow, Quebec, Canada. Geochimica et Cosmochimica Acta 68: 2961-
2980.
Temime-Roussel B; Monod A; Massiani C; Wortham H. (2004a). Evaluation of an annular denuder for atmospheric PAH
partitioning studies—2: evaluation of mass and number particle losses. Atmospheric Environment 38: 1925-1932.
Temime-Roussel B; Monod A; Massiani C; Wortham H. (2004b). Evaluation of an annular denuder tubes for atmospheric PAH
partitioning studies—1: evaluation of the trapping efficiency of gaseous PAHS. Atmospheric Environment 38: 1913-
1924.
ten Brink H; Hoek G; Khlystov A. (2005). An approach to monitor the fraction of elemental carbon in the ultrafine aerosol.
Atmospheric Environment 39: 6255-6259.
ten Brink H; Maenhaut W; Hitzenberger R; Gnauk T; Spindler G; Even A; Chi X; Bauer H; Puxbaum H; Putaud JP. (2004).
INTERCOMP2000: the comparability of methods in use in Europe for measuring the carbon content of aerosol.
Atmospheric Environment 38: 6507-6519.
Tenias JM; Ballester F; Perez-Hoyos S; Rivera ML. (2002). Air pollution and hospital emergency room admissions for chronic
obstructive pulmonary disease in Valencia, Spain. Archives of environmental health 57: 41-47.
Tesche TW; Morris R; Tonnesen G; McNally D; Boylan J; Brewer P. (2006). CMAQ/CAMx annual 2002 performance
evaluation over the eastern US. Atmospheric Environment 40: 4906-4919.
Tesfaigzi Y; McDonald JD; Reed MD; Singh SP; De Sanctis GT; Eynott PR; Hahn FF; Campen MJ; Mauderly JL. (2005). Low-
level subchronic exposure to wood smoke exacerbates inflammatory responses in allergic rats. Toxicological sciences 88:
505-513.
Tesfaigzi Y; Singh SP; Foster JE; Kubatko J; Barr EB; Fine PM; McDonald JD; Hahn FF; Mauderly JL. (2002). Health effects of
subchronic exposure to low levels of wood smoke in rats. Toxicol Sci 65: 115-125.
Textor C, et al. (2006). AeroCom: the status quo of global aerosol modeling. Atmospheric Chemistry and Physics 6: 1777-1813.
Thomas GO; Smith KEC; SweetmanAJ; Jones KC. (1998). Further studies of the air-pasture transfer of polychlorinated
biphenyls. Environmental Pollution 102: 119-128.
Thomson E; Kumarathasan P; Goegan P; Aubin RA; Vincent R. (2005). Differential regulation of the lung endothelin system by
urban particulate matter and ozone. Toxicological sciences 88: 103-113.
Thurlbeck WM. (1982). Postnatal human lung growth. Thorax 37: 564-571.
Thurston GD; Ito K; Mar T; Christensen WF; Eatough DJ; Henry RC; Kim E; Laden F; Lall R; Larson TV. (2005). Workgroup
Report: Workshop on Source Apportionment of Particulate Matter Health Effects—Intercomparison of Results and
Implications. Environmental health perspectives 113: 1768.
Thurston GD; Spengler JD. (1985). A quantitative assessment of source contributions to inhalable particulate matter pollution in
metropolitan Boston. Atmospheric Environment (1967) 19: 9-25.
Tiitanen P; Timonen K; J. Ruskanen J; Mirme A; Pekkanen J. Fine particulate air pollution, resuspended road dust and
respiratory health among symptomatic children. European Respiratory Journal 13: 266-273.
Timmreck C; Shulz M. (2004). Significant dust simulation differences in nudged and climatological operation mode of the
AGCMECHAM. Journal of Geophysical Research 109, D13202, doi:10.1029/2003JD004381.
Timonen KL; Hoek G; Heinrich J; Bernard A; Brunekreef B; de Hartog J; Hameri K; Ibald-Mulli A; Mirme A; Peters A;
Tiittanen P; Kreyling WG; Pekkanen J. (2004). Daily variation in fine and ultrafine particulate air pollution and urinary
concentrations of lung Clara cell protein CC16. Occupational and environmental medicine 61: 908-914.
December 2008
R-69
DRAFT-DO NOT QUOTE OR CITE
-------
Timonen KL; Vanninen E; Hartog ID; Ibald-Mulli A; Brunekreef B; Gold DR; Heinrich J; Hoek G; Lanki T; Peters A;
Tarkiainen T; Tiittanen P; Kreyling W; Pekkanen J. (2006). Effects of ultrafme and fine particulate and gaseous air
pollution on cardiac autonomic control in subjects with coronary artery disease: The ULTRA study. JESEE 16: 332-341.
Tobias HJ; Kooiman PM; Docherty KS; Ziemann PJ. (2000). Real-Time Chemical Analysis of Organic Aerosols Using a
Thermal Desorption Particle Beam Mass Spectrometer. Aerosol Science and Technology 33: 170-190.
Tobias HJ; Ziemann PJ. (1999). Compound Identification in Organic Aerosols Using Temperature-Programmed Thermal
Desorption Particle Beam Mass Spectrometry. ANALYTICAL CHEMISTRY-WASHINGTON DC- 71: 3428-3435.
TobinMJ. (1983). Breathing patterns. 1. Normal subjects. Chest 84: 202-205.
Tolbert PE; Klein M; Metzger KB; Peel J; Flanders WD; Todd K; Mulholland JA; Ryan PB; Frumkin H. (2000). Interim results
of the study of particulates and health in Atlanta (SOPHIA). Journal of exposure analysis and environmental
epidemiology 10: 446-460.
Tolbert PE; Klein M; Peel JL; Sarnat SE; Sarnat JA. (2007). Multipollutant modeling issues in a study of ambient air quality and
emergency department visits in Atlanta. Journal of Exposure Science and Environmental Epidemiology 17: S29-S35.
Tolocka MP; Jang M; Ginter JM; Cox FJ; Kamens RM; Johnston MV. (2004). Formation of Oligomers in Secondary Organic
Aerosol. Environmental science & technology 38: 1428-1434.
Tombach I; McDonald K. (2004). Visibility and Radiative Balance Effects. In P.H. McMurry MFS, and J.S. Vickery (Ed.),
Particulate Matter Science for Policy Makers: A NARSTO Assessment: Cambridge University Press.
Tong C; Blanco M; Goddard WA; Seinfeld JH. (2006). Secondary organic aerosol formation by heterogeneous reactions of
aldehydes and ketones: a quantum mechanical study. Environ Sci Technol 40: 2333-2338.
Tong S; Colditz P. (2004). Air pollution and sudden infant death syndrome: a literature review. Paediatr Perinatal Epidemiol 18:
327-335.
Toose LK; Mackay D. (2004). Adaptation of fugacity models to treat speciating chemicals with constant species concentration
ratios. Environmental Science and Technology 38: 4619^1626.
Tornqvist H; Mills NL; Gonzalez M; Miller MR; Robinson SD; Megson IL; Macnee W; Donaldson K; Soderberg S; Newby DE;
Sandstrom T; Blomberg A. (2007). Persistent endothelial dysfunction in humans after diesel exhaust inhalation.
American journal of respiratory and critical care medicine 176: 395-400.
Tozuka Y; Watanabe N; Osawa M; Toriba A; Kizu R; Hayakawa K. (2004). Transfer of polycyclic aromatic hydrocarbons to
fetuses and breast milk of rats exposed to diesel exhaust. Journal of health science 50: 497-502.
Tran CL; Buchanan D; Cullen RT; Searl A; Jones AD; Donaldson K. (2000). INHALATION OF POORLY SOLUBLE
PARTICLES. II. INFLUENCE OF PARTICLE SURFACE AREA ON INFLAMMATION AND CLEARANCE.
Inhalation Toxicology 12: 1113-1126.
Travnikov O. (2005). Contribution of the intercontinental atmospheric transport to mercury pollution in the Northern
Hemisphere. Atmospheric Environment 39: 7541-7548.
Tremper AH; Agneta M; Burton S; Higgs DEB. (2004). Field and laboratory exposures of two moss species to low level metal
pollution. J Atmos Chem 49: 111-120.
Trenga CA; Sullivan JH; Schildcrout JS; Shepherd KP; Shapiro GG; Liu LJ; Kaufman JD; Koenig JQ. (2006). Effect of
particulate air pollution on lung function in adult and pediatric subjects in a Seattle panel study. Chest 129: 1614-1622.
Trijonis JC; Malm WC; Pitchford M; White WH; Charlson R. (1990). Acidic deposition: State of science and technology. Report
24. Visibility: Existing and historical conditions-causes and effects. Final report. PB-92-100601/XAB, National Acid
Precipitation Assessment Program, Washington, DC (United States),
Tsai CJ; Chang CT; Huang CH. (2006a). Direct field observation of the relative humidity effect on the beta-gauge readings. J Air
Waste Manag Assoc 56: 834-840.
Tsai FC; Apte MG; Daisey JM. (2000). AN EXPLORATORY ANALYSIS OF THE RELATIONSHIP BETWEEN
MORTALITY AND THE CHEMICAL COMPOSITION OF AIRBORNE PARTICULATE MATTER. Inhalation
Toxicology 12: 121-135.
Tsai SS; Chen CC; Hsieh HJ; Chang HH; Yang CY. (2006b). Air pollution and postneonatal mortality in a tropical city:
Kaohsiung, Taiwan. Inhal Toxicol 18: 185-189.
Tsai SS; Cheng MH; Chiu HF; Wu TN; Yang CY. (2006c). Air pollution and hospital admissions for asthma in a tropical city:
Kaohsiung, Taiwan. Inhal Toxicol 18: 549-554.
Tsai SS; Goggins WB; Chiu HF; Yang CY. (2003b). Evidence for an association between air pollution and daily stroke
admissions in Kaohsiung, Taiwan. Stroke; a journal of cerebral circulation 34: 2612-2616.
Tsai SS; Huang CH; Goggins WB; Wu TN; Yang CY. (2003a). Relationship between air pollution and daily mortality in a
tropical city: Kaohsiung, Taiwan. J Toxicol Environ Health A 66: 1341-1349.
Tsai YI; Kuo SC. (2006). Development of diffuse reflectance infrared Fourier transform spectroscopy for the rapid
characterization of aerosols. Atmospheric Environment 40: 1781-1793.
TSI. (2005). Data Merge Software Module for merging and fitting of SMPA and APS Data Files, P/N 1930074, User's Guide,
Model 390069, Revision A. St. Paul, MN. .
Tsukue N; Tsubone H; Suzuki AK. (2002). Diesel exhaust affects the abnormal delivery in pregnant mice and the growth of their
young. Inhal Toxicol 14: 635-651.
Tsukue N; Yoshida S; Sugawara I; Taked K. (2004). Effect of diesel exhaust on development of fetal reproductive function in
ICR female mice. Journal of health science 50: 174-180.
December 2008
R-70
DRAFT-DO NOT QUOTE OR CITE
-------
Tu KW; Knutson EO. (1984). Total Deposition of Ultrafine Hydrophobic and Hygroscopic Aerosols in the Human Respiratory
System. Aerosol Science and Technology 3: 453-465.
Tuch TM; Herbarth O; Franck U; Peters A; Wehner B; Wiedensohler A; Heintzenberg J. (2006). Weak correlation of ultrafine
aerosol particle concentrations< 800 nm between two sites within one city. Journal of Exposure Science and
Environmental Epidemiology 16: 486-491.
Tung TCW; Chao CYH; Burnett J. (1999). A methodology to investigate the particulate penetration coefficient through building
shell. Atmospheric Environment 33: 881-893.
Turpin BJ; Lim HJ. (2001). Species Contributions to PM2. 5 Mass Concentrations: Revisiting Common Assumptions for
Estimating Organic Mass. Aerosol Science and Technology 35: 602-610.
Turpin BJ; Weisel CP; Morandi M; Colome S; Stock T; Eisenreich S; Buckley B. (2007). Relationships of Indoor, Outdoor, and
Personal Air (RIOPA): Part II. Analyses of concentrations of particulate matter species. ResRepHealth Efflnst: 1-77.
U.S. Census Bureau PIO. (2000, Last modified: March 13, 2001). Census Bureau projects doubling of nation's population by
2100. United States Department of Commerce News, from http://www.census.gOv/Press-Release/www/2000/cb00-
05.html
U.S. EPA. (1979). Protecting Visibility, an EPA Report to Congress. EPA-450-5-79-008.
U.S. EPA. (1980). Workbook for Estimating Visibility Impairment. EPA-450-4-80-031.
U.S. EPA. (1982). Air Quality Criteria for Particulate Matter and Sulfur Oxides. U.S. Environmental Protection Agency,
Washington, D.C. EPA/600/8-82/029.
U.S. EPA. (1993). Air Quality Criteria for Oxides of Nitrogen. U.S. Environmental Protection Agency, Office of Research and
Development, National Center for Environmental Assessment, Research Triangle Park, NC.
U.S. EPA. (1996). Air Quality Criteria for Particulate Matter. U.S. Environmental Protection Agency, Office of Research and
Development, National Center for Environmental Assessment, Research Triangle Park, NC. EPA/600/P-95/001aF.
U.S. EPA. (1997). Mercury Study Report to Congress. U.S. Environmental Protection Agency, Office of Research and
Development and Office of Air Quality Planning & Standards EPA-452/R-97-003.
U.S. EPA. (1998). SLAMS / NAMS / PAMS network review guidance. Available:
http://www.epa.gov/ttn/amtic/files/ambient/criteria/reldocs/netrev98.pdf (22 February, 2008). U.S. Environmental
Protection Agency, Office of Air and Radiation, Office of Air Quality Planning and Standards, Research Triangle Park,
NC. EPA-454/R-98-003.
U. S. EPA. (2001). Draft Guidance for tracking progress under the regional haze rule. EPA-450-4-80-031.
http://vista.cira.colostate.edu/improve/Publications/GuidanceDocs/guidancedocs.htm.
U.S. EPA. (2004). Air Quality Criteria for Particulate Matter. U.S. Environmental Protection Agency, Washington, D.C.
EPA/600/P-99/002aF-bF.
U.S. EPA. (2005a). Guidelines for Carcinogen Risk Assessment. U.S. Environmental Protection Agency, Washington, D.C.
EPA/630/P-03/001F.
U.S. EPA. (2005b). Review of the National Ambient Air Quality Standards for Particulate Matter: Policy Assessment of
Scientific and Technical Information. OAQPS Staff Paper. U.S. Environmental Protection Agency, Office of Air and
Radiation, Office of Air Quality Planning and Standards, Research Triangle Park, NC. EPA-452/R-05-005a.
U.S. EPA. (2006a). 2002 National Emissions Inventory Data and Documentation. U.S. Environmental Protection Agency, Office
of Air and Radiation, Office of Air Quality Planning and Standards, Research Triangle Park, NC.
U.S. EPA. (2006b). Air Quality Criteria for Lead. U.S. Environmental Protection Agency, Office of Research and Development,
National Center for Environmental Assessment, Research Triangle Park, NC.
U.S. EPA. (2006c). Air Quality Criteria for Ozone and Related Photochemical Oxidants. U.S. Environmental Protection Agency,
Office of Research and Development, National Center for Environmental Assessment, Research Triangle Park, NC. EPA
600/R-05/004aF.
U.S. EPA. (2006d). Provisional Assessment of Recent Studies on Health Effects of Particulate Matter Exposure. U.S.
Environmental Protection Agency, Office of Research and Development, National Center for Environmental Assessment,
Washington, DC. EPA 600/R-06/063.
U.S. EPA. (2008a). U.S. EPA's 2008 Report on the Environment (Final Report). U.S. Environmental Protection Agency, Office
of Research and Development, National Center for Environmental Assessment, Washington, DC. EPA/600/R-07/045F.
U.S. EPA. (2008b). Integrated Review Plan for the National Ambient Air Quality Standards for Particulate Matter. U.S.
Environmental Protection Agency, Office of Research and Development, National Center for Environmental Assessment,
Research Triangle Park, NC.
U.S. EPA. (2008c). Integrated Science Assessment for Oxides of Nitrogen - Health Criteria. U.S. Environmental Protection
Agency, Office of Research and Development, National Center for Environmental Assessment, Research Triangle Park,
NC. EPA/600/R-08/071.
U.S. EPA. (2008d). Integrated Science Assessment for Sulfur Oxides-Health Criteria. U.S. Environmental Protection Agency,
Office of Research and Development, National Center for Environmental Assessment, Research Triangle Park, NC.
EPA/600/R-08/047F-FA.
U.S. EPA. (2008e). Integrated Science Assessment for Oxides of Nitrogen and Sulfur—Ecological Criteria. U.S. Environmental
Protection Agency, Office of Research and Development, National Center for Environmental Assessment, Research
Triangle Park, NC. EPA/600/R-08/082F.
December 2008 R-71 DRAFT-DO NOT QUOTE OR CITE
-------
Ulirsch GV; Ball LM; Kaye W; Shy CM; Lee CV; Crawford-Brown D; Symons M; Holloway T. (2007). Effect of particulate
matter air pollution on hospital admissions and medical visits for lung and heart disease in two southeast Idaho cities.
Journal of Exposure Science and Environmental Epidemiology 17: 478-487.
UNCEC (2007). Hemispheric Transport of Air Pollution 2007. In Task Force on Hemispheric Transport of Air Pollution (Eds.),
Available from http://www.htap.org/activities/2007_Interim_Report.htm
Unsworth MH; Wilshaw JC. (1989). Wet, occult and dry deposition of pollutants on forests. Agricultural and Forest Meteorology
47: 221-238.
Upadhyay D; Panduri V; Ghio A; Kamp DW. (2003). Particulate matter induces alveolar epithelial cell DNA damage and
apoptosis: role of free radicals and the mitochondria. American journal of respiratory cell and molecular biology 29: 180-
187.
Urch B; Brook JR; Wasserstein D; Brook RD; Rajagopalan S; Corey P; Silverman F. (2004). Relative contributions of PM2.5
chemical constituents to acute arterial vasoconstriction in humans. Inhal Toxicol 16: 345-352.
Urch B; Silverman F; Corey P; Brook JR; Lukic KZ; Rajagopalan S; Brook RD. (2005). Acute blood pressure responses in
healthy adults during controlled air pollution exposures. Environ Health Perspect 113: 1052-1055.
Utgikar VP; Chaudhary N; Koeniger A; Tabak HH; Haines JR; Govind R. (2004). Toxicity of metals and metal mixtures:
Analysis of concentration and time dependence for zinc and copper. Water Research 38: 3651-3658.
Vaisvalavicius R; Motuzas A; Prosycevas I; Levinskaite L; Zakarauskaite D; Grigaliuniene K; Butkus V. (2006). Effect of heavy
metals on microbial communities and enzymatic activity in soil column experiment. Archives of Agronomy and Soil
Science 52: 161-169.
Valavanidis A; Vlahoyianni T; Fiotakis K. (2005). Comparative study of the formation of oxidative damage marker 8-hydroxy-
2'-deoxyguanosine (8-OHdG) adduct from the nucleoside 2'-deoxyguanosine by transition metals and suspensions of
particulate matter in relation to metal content and redox reactivity. Free radical research 39: 1071-1081.
Vallejo M; Ruiz S; Hermosillo AG; Boija-Aburto VH; Cardenas M. (2006). Ambient fine particles modify heart rate variability
in young healthy adults. Journal of Exposure Science and Environmental Epidemiology 16: 125-130.
Van Aalst RM. (1982). Dry deposition ofNOx. In Schneider T, Grant L (Eds.), Air pollution by nitrogen oxides (pp. 263-270).
Amsterdam, The Netherlands: Elsevier Scientific Publishing Company,
van de Hulst HC. (1981). Light Scattering by Small Particles: Courier Dover Publications.
van den Broek I; Sparidans RW; Schellens JH; Beijnen JH. (2008). Liquid chromatography/tandem mass spectrometric method
for the quantification of eight proteolytic fragments of ITIH(4) with biomarker potential in human plasma and serum.
Rapid Commun Mass Spectrom 22: 2915-2928.
van der Meer IM; Bots ML; Hofman A; del Sol AI; van der Kuip DA; Witteman JC. (2004). Predictive value of noninvasive
measures of atherosclerosis for incident myocardial infarction: the Rotterdam Study. Circulation 109: 1089-1094.
van der Werf GR; Randerson JT; Giglio L; Collatz GJ; Kasibhatla PS; Arellano Jr AF. (2006). Interannual variability in global
biomass burning emissions from 1997 to 2004. Atmos Chem Phys 6: 3423-3441.
van Dingenen R; Raes F; Putaud JP; Baltensperger U; Charron A; Facchini MC; Decesari S; Fuzzi S; Gehrig R; Hansson HC;
Harrison RM; Hiiglin C; Jones AM; Laj P; Lorbeer G; Maenhaunt W; Palmgren F; Querol X; Rodriguez S; Schneider J;
ten Brink H; Tunved P; T0rseth K; Wehner B; Weingartner E; Wiedensohler A; Wahlin P. (2004). A European aerosol
phenomenology - 1: physical characteristics of particulate matter at kerbside, urban, rural and background sites in Europe.
Atmospheric Environment 38: 2561-2577.
van Eeden SF; Hogg JC. (2002). Systemic Inflammatory Response Induced by Particulate Matter Air Pollution: The Importance
of Bone-Marrow Stimulation. Journal of Toxicology and Environmental Health, Part A 65: 1597-1613.
van Eeden SF; Tan W; Suwa T; Mukae H; Terashima T; Fujii T; Qui D; Vincent R; Hogg JC. (2001). Cytokines Involved in the
Systemic Inflammatory Response Induced by Exposure to Particulate Matter Air Pollutants (PM10). American journal of
respiratory and critical care medicine 164: 826-830.
van Eeden SF; Yeung A; Quinlam K; Hogg JC. (2005). Systemic response to ambient particulate matter: relevance to chronic
obstructive pulmonary disease. Proceedings of the American Thoracic Society 2: 61-67.
van Heerden PDR; Kriiger GHJ; Kilbourn Louw M. (2007). Dynamic responses of photosystem II in the Namib Desert shrub,
Zygophyllum prismatocarpum, during and after foliar deposition of limestone dust. Environmental Pollution 146: 34-45.
Van Metre PC; Mahler BJ. (2003). The contribution of particles washed from rooftops to contaminant loading to urban streams.
Chemosphere 52: 1727-1741.
VanCuren RA; Cahill TA. (2002). Asian aerosols in North America: Frequency and concentration of fine dust. Journal of
Geophysical Research D: Atmospheres 107 (2002) pArticle Number: 4804: Article.
Vastag E; Matthys H; Kohler D; Gronbeck L; Daikeler G. (1985). Mucociliary clearance and airways obstruction in smokers, ex-
smokers and normal subjects who never smoked. Eur J Respir Dis Suppl 139: 93-100.
Vasundhara G; Jayashree G; Muraleedhara-Kurup G. (2004). Sequestration of nickel and copper by Azotobacter chroococcum
SB1. Bulletin of Environmental Contamination and Toxicology 72: 1122-1127.
Vedal S; Rich K; Brauer M; White R; Petkau J. (2004). Air pollution and cardiac arrhythmias in patients with implantable
cardioverter defibrillators. Inhal Toxicol 16: 353-362.
Vega E; Reyes E; Wellens A; Sanchez G; Chow JC; Watson JG. (2003). Comparison of continuous and filter based mass
measurements in Mexico City. Atmospheric Environment 37: 2783-2793.
Vegni FE; Ros O. (2004). Hospital Accident and Emergency burden is unaffected by today's air pollution levels. Eur J Emerg
Med 11: 86-88.
December 2008
R-72
DRAFT-DO NOT QUOTE OR CITE
-------
Venkataraman C; Lyons JM; Friedlander SK. (1994). Size Distributions of Poly cyclic Aromatic Hydrocarbons and Elemental
Carbon. 1. Sampling, Measurement Methods, and Source Characterization. Environmental science & technology 28: 555-
562.
Veranth JM; Moss TA; Chow JC; Labban R; Nichols WK; Walton JC; Watson JG; Yost GS. (2006). Correlation of in vitro
cytokine responses with the chemical composition of soil-derived particulate matter. Environmental health perspectives
114: 341-349.
Veronesi B; Makwana O; Pooler M; Chen Lung C. (2005). Effects of subchronic exposures to concentrated ambient particles.
VII. Degeneration of dopaminergic neurons in Apo E-/- mice. Inhalation toxicology 17: 235-241.
Veronesi B; Oortgiesen M. (2001). Neurogenic Inflammation and Particulate Matter (PM) Air Pollutants. Neurotoxicology 22:
795-810.
Veronesi B; Oortgiesen M; Carter JD; Devlin RB. (1999). Particulate Matter Initiates Inflammatory Cytokine Release by
Activation of Capsaicin and Acid Receptors in a Human Bronchial Epithelial Cell Line. Toxicology and applied
pharmacology 154: 106-115.
Veronesi B; Oortgiesen M; Roy J; Carter JD; Simon SA; Gavett SH. (2000). Vanilloid (Capsaicin) Receptors Influence
Inflammatory Sensitivity in Response to Particulate Matter. Toxicology and applied pharmacology 169: 66-76.
Veronesi B; Wei G; Zeng JQ; Oortgiesen M. (2003). Electrostatic Charge Activates Inflammatory Vanilloid (VR1) Receptors.
Neurotoxicology 24: 463-473.
Viana M; Querol X; Alastuey A. (2006). Chemical characterisation of PM episodes in NE Spain. Chemosphere 62: 947-956.
Viana M; Querol X; Alastuey A; Ballester F; Llop S; Esplugues A; FernAjndez-Patier R; GarcA-a dos Santos S; Herce MD.
(2008). Characterising exposure to PM aerosols for an epidemiological study. Atmospheric Environment 42: 1552-1568.
Viard B; Pihan F; Promeyrat S; Pihan J-C. (2004). Integrated assessment of heavy metal (Pb, Zn, Cd) highway pollution:
bioaccumulation in soil, Graminaceae and land snails. Chemosphere 55: 1349-1359.
Vigotti MA; Chiaverini F; Biagiola P; Rossi G. (2007). Urban air pollution and emergency visits for respiratory complaints in
Pisa, Italy. J Toxicol Environ Health A 70: 266-269.
Viles HA; Gorbushina AA. (2003). Soiling and microbial colonisation on urban roadside limestone: A three year study in
Oxford, England. Building and Environment 38: 1217-1224.
Viles HA; Taylor MP; Yates TJS; Massey SW. (2002). Soiling and decay of N.M.E.P. limestone tablets. Science of the Total
Environment 292: 215-229.
Villa S; Vighi M; Maggi V; Finizio A; Bolzacchini E. (2003). Historical trends of organochlorine pesticides in an Alpine glacier.
JAtmosChem46: 295-311.
Villeneuve PJ, Goldberg MS, Krewski D, Burnett RT, Chen Y. 2002. Fine particulate air pollution and all-cause
mortality within the Harvard Six-Cities Study: variations in risk by period of exposure. Ann Epidemiol 12(8):
568-576.
Villeneuve PJ; Burnett RT; Shi Y; Krewski D; Goldberg MS; Hertzman C; Chen Y; Brook J. (2003). A time-series study of air
pollution, socioeconomic status, and mortality in Vancouver, Canada. Journal of exposure analysis and environmental
epidemiology 13: 427-435.
Villeneuve PJ; Chen L; Stieb D; Rowe BH. (2006). Associations between outdoor air pollution and emergency department visits
for stroke in Edmonton, Canada. European journal of epidemiology 21: 689-700.
Villeneuve PJ; Goldberg MS; Krewski D; Burnett RT; Chen Y. (2002). Fine particulate air pollution and all-cause mortality
within the Harvard Six-Cities Study: variations in risk by period of exposure. Ann Epidemiol 12: 568-576.
Vincent JH; Donaldson K. (1990). A dosimetric approach for relating the biological response of the lung to the accumulation of
inhaled mineral dust. British Journal of Industrial Medicine 47: 302.
Vincent R; Kumarathasan P; Goegan P; Bjarnason SG; Guenette J; Berube D; Adamson IY; Desjardins S; Burnett RT; Miller FJ;
Battistini B. (2001). Inhalation toxicology of urban ambient particulate matter: acute cardiovascular effects in rats. Res
Rep Health Eff Inst: 5-54; discussion 55-62.
Vinzents PS; Moller P; Sorensen M; Knudsen LE; Hertel O; Jensen FP; Schibye B; Loft S. (2005). Personal exposure to ultrafme
particles and oxidative DNA damage. Environ Health Perspect 113: 1485-1490.
Violante FS; Barbieri A; Curti S; Sanguinetti G; Graziosi F; Mattioli S. (2006). Urban atmospheric pollution: Personal exposure
versus fixed monitoring station measurements. Chemosphere 64: 1722-1729.
Virkkula A; Makela T; Hillamo R; Yli-Tuomi T; Hirsikko A; Hameri K; Koponen IK. (2007). A simple procedure for correcting
loading effects of aethalometer data. J Air Waste ManagAssoc 57: 1214-1222.
Viskari E-L; Karenlampi L. (2000). Roadside Scots pine as an indicator of deicing salt use - a comparative study from two
consecutive winters. Water, Air, & Soil Pollution 122: 405-419.
Vives I; Grimalt JO; Ventura M; Catalan J. (2005). Distribution of polycyclic aromatic hydrocarbons in the food web of a high
mountain lake, Pyrenees, Catalonia, Spain. Environmental Toxicology & Chemistry 24: 1344-1352.
VogtMT; Cauley JA; Newman AB; KullerLH; Hulley SB. (1993). Decreased ankle/arm blood pressure index and mortality in
elderly women. JAMA 270: 465-469.
Voisin D; Smith JN; Sakurai H; McMurry PH; Eisele FL. (2003). Thermal Desorption Chemical Ionization Mass Spectrometer
for Ultrafme Particle Chemical Composition. Aerosol Science and Technology 37: 471-475.
von Klot S; Peters A; Aalto P; Bellander T; Berglind N; DTppoliti D; Elosua R; Hormann A; Kulmala M; Lanki T; Lowel H;
Pekkanen J; Picciotto S; Sunyer J; Forastiere F. (2005). Ambient air pollution is associated with increased risk of hospital
cardiac readmissions of myocardial infarction survivors in five European cities. Circulation 112: 3073-3079.
December 2008
R-73
DRAFT-DO NOT QUOTE OR CITE
-------
von Klot S; Wolke G; Tuch T; Heinrich J; Dockery DW; Schwartz J; Kreyling WG; Wichmann HE; Peters A. (2002). Increased
asthma medication use in association with ambient fine and ultrafrne particles. Eur Respir J 20: 691-702.
Vyas VM; Christakos G. (1997). Spatiotemporal analysis and mapping of sulfate deposition data over Eastern USA. Atmospheric
Environment 31: 3623-3633.
Wagner EM; Foster WM. (2001). Interdependence of bronchial circulation and clearance of 99mTc-DTPA from the airway
surface. Journal of applied physiology 90: 1275-1281.
Wallace L. (2000). Real-Time Monitoring of Particles, PAH, and CO in an Occupied Townhouse. Applied occupational and
environmental hygiene 15: 39-47.
Wallace L. (2005). Ultrafrne particles from a vented gas clothes dryer. Atmospheric Environment 39: 5777-5786.
Wallace L; Croghan C; Williams R; Rea A. (2006). Continuous weeklong measurements of personal exposures and indoor
concentrations of fine particles for 37 health-impaired North Carolina residents for up to four seasons. Atmospheric
Environment 40: 399-414.
Wallace L; Williams R. (2005). Use of personal-indoor-outdoor sulfur concentrations to estimate the infiltration factor and
outdoor exposure factor for individual homes and persons. EnvironSci Technol 39: 1707-1714.
Wallenborn JG; McGee John K; Schladweiler Mette C; Ledbetter Allen D; Kodavanti Urmila P. (2007). Systemic translocation
of particulate matter-associated metals following a single intratracheal instillation in rats. Toxicological sciences 98: 231-
239.
Walsh CR; Cupples LA; Levy D; Kiel DP; Hannan M; Wilson PW; O'Donnell CJ. (2002). Abdominal aortic calcific deposits are
associated with increased risk for congestive heart failure: the Framingham Heart Study. Am Heart J 144: 733-739.
Wan J; Diaz-Sanchez D. (2006). Phase II enzymes induction blocks the enhanced IgE production in B cells by diesel exhaust
particles. Journal of immunology 177: 3477-3483.
Wan J; Diaz-Sanchez D. (2007). Antioxidant enzyme induction: a new protective approach against the adverse effects of diesel
exhaust particles. Inhalation toxicology 19 Suppl 1: 177-182.
Wang J; Christopher SA; Nair US; Reid JS; Prins EM; Szykman J; Hand JL. (2006a). Mesoscale modeling of Central American
smoke transport to the United States: 1. "Top-down"assessment of emission strength and diurnal variation impacts. J
GeophysRes 111.
Wang JX; Chen CY; Yu HW; Sun J; Li B; Li YF; Gao YX; He W; Huang YY; Chai ZF. (2007). Distribution of TiO 2 particles
in the olfactory bulb of mice after nasal inhalation using microbeam SRXRF mapping techniques. Journal of
Radioanalytical and Nuclear Chemistry 272: 527-531.
Wang X; Bi X; Sheng G; Fu J. (2006b). Chemical composition and sources of PM10 and PM2.5 aerosols in Guangzhou, China.
Environ Monit Assess 119: 425-439.
Wang X; Ding H; Ryan L; Xu X. (1997). Association between air pollution and low birth weight: a community-based study.
Environ Health Perspect 105: 514-520.
Wang Y-Z; Ingram JL; Walters DM; Rice AB; Santos JH; Van Houten B; Bonner JC. (2003). Vanadium-induced STAT-1
activation in lung myofibroblasts requires H202 and P38 MAP kinase. Free radical biology & medicine 35: 845-855.
Wania F; Hoff JT; Jia CQ; Mackay D. (1998). The effects of snow and ice on the environmental behaviour of hydrophobic
organic chemicals. Environmental Pollution 102: 43-51.
Wania F; Mackay D. (1993). Global fractionation and cold condensation of low volatility organochlorine compounds in polar
regions. Ambio 22: 10-18.
Ward PA. (2003). Acute lung injury: how the lung inflammatory response works. Eur Respir J Suppl 44: 22s-23s.
Warheit DB; Hansen JF; Yuen IS; Kelly DP; Snajdr SI; Hartsky MA. (1997). Inhalation of High Concentrations of Low Toxicity
Dusts in Rats Results in Impaired Pulmonary Clearance Mechanisms and Persistent Inflammation. Toxicology and
applied pharmacology 145: 10-22.
Warheit DB; Webb TR; Colvin VL; Reed KL; Sayes CM. (2007). Pulmonary Bioassay Studies with Nanoscale and Fine-Quartz
Particles in Rats: Toxicity is Not Dependent upon Particle Size but on Surface Characteristics. Toxicological sciences 95:
270.
Warheit DB; Webb TR; Sayes CM; Colvin VL; Reed KL. (2006). Pulmonary instillation studies with nanoscale Ti02 rods and
dots in rats: toxicity is not dependent upon particle size and surface area. Toxicol Sci 91: 227-236.
Warner J. (1968). A reduction in rainfall associated with smoke from sugar-cane fires—an inadvertent weather modification?
Journal of Applied Meteorology 7: 247-251.
Warner J; Twomey SA. (1967). The production and cloud nuclei by cane fires and the effect on cloud droplet concentration.
Journal of Atmospheric Science 24: 704-706.
Watanabe N. (2005). Decreased number of sperms and Sertoli cells in mature rats exposed to diesel exhaust as fetuses.
Toxicology letters 155: 51-58.
Watanabe N; Tanada S; Sasaki Y. (2007). Pulmonary clearance of aerosolized 99mTc-DTPA in sarcoidosis I patients. QJ Nucl
Med Mol Imaging 51: 82-90.
Watkinson WP; Campen MJ; Costa DL. (1998). Cardiac arrhythmia induction after exposure to residual oil fly ash particles in a
rodent model of pulmonary hypertension. Toxicol Sci 41: 209-216.
Watmough SA; Hutchinson TC; Dillon PJ. (2004). Lead dynamics in the forest floor and mineral soil in south-central Ontario.
Biogeochemistry 71: 43-68.
Watson GJ; Farrell P; Stanton S; Skidmore LC. (2007). Effects of Bait Collection On Nereis Virens Populations and Macrofaunal
Communities in the Solent, Uk. J Mar Biol Assoc UK 87: 703-716.
December 2008
R-74
DRAFT-DO NOT QUOTE OR CITE
-------
Watson JG; Chen LW; Chow JC; Doraiswamy P; Lowenthal DH. (2008). Source apportionment: findings from the U.S.
Supersites Program. J Air Waste Manag Assoc 58: 265-288.
Watson JG; Chow JC. (2002). Comparison and evaluation of in situ and filter carbon measurements at the Fresno Supersite. J
GeophysRes 107: 3-1.
Watson JG; Chow JC. (2007). RECEPTOR MODELS FOR SOURCE APPORTIONMENT OF SUSPENDED PARTICLES.
Introduction to Environmental Forensics.
Watson JG; Chow JC; Chen LWA. (2005b). Summary of organic and elemental carbon/black carbon analysis methods and
intercomparisons. Aerosol and Air Quality Research 5: 69-102.
Watson JG; Chow JC; Doraiswamy P; Chen LWA; Lowenthal DH; Trimble D; Park K. (2005a). Quality Assurance Final Report
for the Fresno Supersite. Desert Research Institute, Reno, NV.
Watson JG; Chow JC; DuBois D; Green M; Frank N; Pitchford M. (1997). Guidance for Network Design and Optimum Site
Exposure for PM 2.5 and PM 10. PB—99-157513/XAB, Environmental Protection Agency, Office of Air Quality
Planning and Standards, Research Triangle Park, NC (United States); Nevada Univ. System, Desert Research Inst., Reno,
NV (United States); National Oceanic and Atmospheric Administration, Las Vegas, NV (United States),
Watson JG; Office of Vice President for R; Information T; Desert Research I; Specialists IAR; University of Nevada S; Inc ST;
Colorado State U. (1998). Northern Front Range Air Quality Study Final Report: Colorado State University, Office of the
Vice President for Research and Information Technology.
Watson JG; Robinson NF; Chow JC; Henry RC; Kim BM; Pace TG; Meyer EI; Nguyen Q. (1990). The USEPA/DRI chemical
mass balance receptor model. Envir Software 5: 38-49.
Watson JG; Zhu T; Chow JC; Engelbrecht J; Fujita EM; Wilson WE. (2002). Receptor modeling application framework for
particle source apportionment. Chemosphere 49: 1093-1136.
Weber P; Reznicek L; Mitteregger G; Kretzschmar H; Giese A. (2008). Differential effects of prion particle size on infectivity in
vivo and in vitro. Biochem Biophys Res Commun 369: 924-928.
Weber RJ; Orsini D; Daun Y; Lee YN; Klotz PJ; Brechtel F. (2001). A Particle-into-Liquid Collector for Rapid Measurement of
Aerosol Bulk Chemical Composition. Aerosol Science and Technology 35: 718-727.
Weber RJ; Orsini D; Duan Y; Baumann K; Kiang CS; Chameides W; Lee YN; Brechtel F; Klotz P; Jongejan P. (2003).
Intercomparison of near real time monitors of PM2. 5 nitrate and sulfate at the US Environmental Protection Agency
Atlanta Supersite. J Geophys Res 108: 8421.
Weijers EP; Khlystov AY; Kos GPA; Erisman JW. (2004). Variability of particulate matter concentrations along roads and
motorways determined by a moving measurement unit. Atmospheric Environment 38: 2993-3002.
Weingartner E; SaathoffH; SchnaiterM; StreitN; BitnarB; Baltensperger U. (2003). Absorption of light by soot particles:
determination of the absorption coefficient by means of aethalometers. J Aerosol Sci 34: 1445-1463.
Weitz JI; Byrne J; Clagett GP; Farkouh ME; Porter JM; Sackett DL; Strandness DE, Jr.; Taylor LM. (1996). Diagnosis and
treatment of chronic arterial insufficiency of the lower extremities: a critical review. Circulation 94: 3026-3049.
Welin L; Eriksson H; Larsson B; Svardsudd K; Wilhelmsen L; Tibblin G. (1993). Risk Factors for Coronary Heart Disease
during 25 Years of Follow-Up. Cardiology 82: 223-228.
Wellenius GA; Batalha JRF; Diaz EA; Lawrence J; Coull BA; Katz T; Verrier RL; Godleski JJ. (2004). Cardiac effects of carbon
monoxide and ambient particles in a rat model of myocardial infarction. Toxicological sciences 80: 367-376.
Wellenius GA; Bateson TF; Mittleman MA; Schwartz J. (2005b). Particulate air pollution and the rate of hospitalization for
congestive heart failure among medicare beneficiaries in Pittsburgh, Pennsylvania. Am J Epidemiol 161: 1030-1036.
Wellenius GA; Coull BA; Batalha JRF; Diaz EA; Lawrence J; Godleski JJ. (2006a). Effects of ambient particles and carbon
monoxide on supraventricular arrhythmias in a rat model of myocardial infarction. Inhalation Toxicology 18: 1077-1082.
Wellenius GA; Coull BA; Godleski JJ; Koutrakis P; Okabe K; Savage ST; Lawrence JE; Murthy GG; Verrier RL. (2003).
Inhalation of concentrated ambient air particles exacerbates myocardial ischemia in conscious dogs. Environ Health
Perspect 111: 402-408.
Wellenius GA; Saldiva PH; Batalha JR; Krishna Murthy GG; Coull BA; Verrier RL; Godleski JJ. (2002). Electrocardiographic
changes during exposure to residual oil fly ash (ROFA) particles in a rat model of myocardial infarction. Toxicol Sci 66:
327-335.
Wellenius GA; Schwartz J; Mittleman MA. (2005a). Air pollution and hospital admissions for ischemic and hemorrhagic stroke
among medicare beneficiaries. Stroke; a journal of cerebral circulation 36: 2549-2553.
Wellenius GA; Schwartz J; Mittleman MA. (2006b). Particulate air pollution and hospital admissions for congestive heart failure
in seven United States cities. Am J Cardiol 97: 404-408.
Wellenius GA; Yeh GY; Coull BA; Suh HH; Phillips RS; Mittleman MA. (2007). Effects of ambient air pollution on functional
status in patients with chronic congestive heart failure: a repeated-measures study. Environ Health 6: 26.
Welton EJ; Campbell JR; Spinhirne JD; Scott VS. (2001). Global monitoring of clouds and aerosols using a network of micro-
pulse lidar systems. In Singh UN, Itabe T, Sugimoto N (Eds.), Lidar remote sensing for industry and environmental
monitoring (pp. 151-158). Bellingham, WA: SPIE.
Welty LJ; Zeger SL. (2005). Are the acute effects of particulate matter on mortality in the National Morbidity, Mortality, and Air
Pollution Study the result of inadequate control for weather and season? A sensitivity analysis using flexible distributed
lag models. Am J Epidemiol 162: 80-88.
Wendelhag I; Gustavsson T; Suurkula M; Berglund G; Wikstrand J. (1991). Ultrasound measurement of wall thickness in the
carotid artery: fundamental principles and description of a computerized analysing system. Clin Physiol 11: 565-577.
December 2008
R-75
DRAFT-DO NOT QUOTE OR CITE
-------
Wendelhag I; Wiklund O; Wikstrand J. (1993). Atherosclerotic changes in the femoral and carotid arteries in familial
hypercholesterolemia. Ultrasonographic assessment of intima-media thickness and plaque occurrence. Arterioscler
Thromb 13: 1404-1411.
Wenzel RJ; Liu DY; Edgerton ES; Prather KA. (2003). Aerosol time-of-flight mass spectrometry during the Atlanta Supersite
Experiment: 2. Scaling procedures. Journal of Geophysical Research (Atmospheres 108.
Wesely ML; Eastman JA; Stedman DH; Yalvac ED. (1982). An eddy-correlation measurement of N02 flux to vegetation and
comparison to 03 flux. Atmospheric Environment 16: 815-820.
Westerdahl D; Fruin S; Sax T; Fine PM; Sioutas C. (2005). Mobile platform measurements of ultrafine particles and associated
pollutant concentrations on freeways and residential streets in Los Angeles. Atmospheric Environment 39: 3597-3610.
Weynand B; Jonckheere A; Frans A; Rahier J; Saint-Luc CU. (1999). Diabetes mellitus Induces a Thickening of the Pulmonary
Basal Lamina. Logo 66.
Whaley SL; Muggenburg BA; Seiler FA; Wolff RK. (1987). Effect of aging on tracheal mucociliary clearance in beagle dogs.
Journal of applied physiology 62: 1331-1334.
Wheeler A; Zanobetti A; Gold DR; Schwartz J; Stone P; Suh HH. (2006a). The relationship between ambient air pollution and
heart rate variability differs for individuals with heart and pulmonary disease. Environ Health Perspect 114: 560-566.
Wheeler EF; Diehl NK; Zajaczkowski J; Brown D. (2006b). Particulate matter characterization in equestrian riding arenas.
Transactions of the ASABE 49: 1529-1538.
Whitby KT. (1978). The physical characteristics of sulfur aerosols. Atmospheric Environment (1967) 12: 135-159.
Whitekus MJ; Li N; Zhang M; Wang M; Horwitz MA; Nelson SK; Horwitz LD; Brechun N; Diaz-Sanchez D; Nel AE. (2002).
Thiol antioxidants inhibit the adjuvant effects of aerosolized diesel exhaust particles in a murine model for ovalbumin
sensitization. J Immunol 168: 2560-2567.
Wichmann HE; Spix C; Tuch T; Wolke G; Peters A; Heinrich J; Kreyling WG; Heyder J. (2000). Daily mortality and fine and
ultrafine particles in Erfurt, Germany part I: role of particle number and particle mass. Res Rep Health Eff Inst 98: 5-86.
Widdicombe J. (2006). Reflexes from the lungs and airways: historical perspective. Journal of applied physiology 101: 628.
Widdicombe J; Lee LY. (2001). Airway Reflexes, Autonomic Function, and Cardiovascular Responses. Environmental health
perspectives 109: 579-584.
Widdicombe JG. (2003). Overview of neural pathways in allergy and asthma. Pulmonary Pharmacology & Therapeutics 16: 23-
30.
Wiebert P; Sanchez-Crespo A; Falk R; Philipson K; Lundin A; Larsson S; Moller W; Kreyling W; Svartengren M. (2006a). No
Significant Translocation of Inhaled 35-nm Carbon Particles to the Circulation in Humans. Inhalation Toxicology 18:
741-747.
Wiebert P; Sanchez-Crespo A; Seitz J; Falk R; Philipson K; Kreyling WG; Moller W; Sommerer K; Larsson S; Svartengren M.
(2006b). Negligible clearance of ultrafine particles retained in healthy and affected human lungs. Eur Respir J 28: 286-
290.
Wild E; Dent J; Thomas GO; Jones KC. (2005). Direct observation of organic contaminant uptake, storage, and metabolism
within plant roots. Environmental Science and Technology 39: 3695-3702.
Wild SR; Jones KC. (1992). Polynuclear aromatic hydrocarbons uptake by carrots grown in sludge amended soil. J Environ Qual
21: 217-225.
Wilhelm M; Ritz B. (2005). Local variations in CO and particulate air pollution and adverse birth outcomes in Los Angeles
County, California, USA. Environ Health Perspect 113: 1212-1221.
Willekes C; Brands PJ; Willigers JM; Hoeks AP; Reneman RS. (1999). Assessment of local differences in intima-media
thickness in the human common carotid artery. J Vase Res 36: 222-228.
Williams B; Goldstein A; Kreisberg N; Hering S. (2006a). An In-Situ Instrument for Speciated Organic Composition of
Atmospheric Aerosols: Thermal Desorption Aerosol GC/MS-FID (TAG). Aerosol Science and Technology 40: 627-638.
Williams R; Case M; Yeatts K; Chen F-L; Scott J; Svendsen E; Devlin R. (2008). Personal coarse particulate matter exposures in
an adult cohort. Atmospheric Environment 42: 6743-6748.
Williams TC; Shaddix CR; Jensen KA; Suo-Anttila JM. (2006b). Measurements oof the Dimensionless Extinction Coefficient of
Soot Within Laminar Diffusion Flames. Submitted to Combustion and Flame.
Wilson JG; Zawar-Reza P. (2006). Intraurban-scale dispersion modelling of particulate matter concentrations: Applications for
exposure estimates in cohort studies. Atmospheric Environment 40: 1053-1063.
Wilson MR; Foucaud L; Barlow PG; Hutchison GR; Sales J; Simpson RJ; Stone V. (2007a). Nanoparticle interactions with zinc
and iron: Implications for toxicology and inflammation. Toxicology and applied pharmacology 225: 80-89.
Wilson PW; Kauppila LI; O'Donnell CJ; Kiel DP; Hannan M; Polak JM; Cupples LA. (2001). Abdominal aortic calcific deposits
are an important predictor of vascular morbidity and mortality. Circulation 103: 1529-1534.
Wilson WE; Brauer M. (2006). Estimation of ambient and non-ambient components of particulate matter exposure from a
personal monitoring panel study. J ExpoSci EnvironEpidemiol 16: 264-274.
Wilson WE; Grover BD; Long RW; Eatough NL; Eatough DJ. (2006). The measurement of fine particulate semivolatile material
in urban aerosols. J Air Waste Manag Assoc 56: 384-397.
Wilson WE; Mage DT; Grant LD. (2000). Estimating separately personal exposure to ambient and nonambient particulate matter
for epidemiology and risk assessment: why and how. Journal of the Air & Waste Management Association 50: 1167-
1183.
December 2008
R-76
DRAFT-DO NOT QUOTE OR CITE
-------
Wilson WE; Mar TF; Koenig JQ. (2007b). Influence of exposure error and effect modification by socioeconomic status on the
association of acute cardiovascular mortality with particulate matter in Phoenix. Journal of Exposure Science and
Environmental Epidemiology 17: SI 1.
Wilson WE; Suh HH. (1997). Fine particles and coarse particles: Concentration relationships relevant to epidemiologic studies.
Journal of the Air & Waste Management Association 47: 1238-1249.
Winkler PM; Steiner G; Vrtala A; Vehkamaki H; Noppel M; Lehtinen KEJ; Reischl GP; Wagner PE; Kulmala M. (2008).
Eleterogeneous Nucleation Experiments Bridging the Scale from Molecular Ion Clusters to Nanoparticles. Science 319:
1374.
Witteman JC; Kok FJ; van Saase JL; Valkenburg UA. (1986). Aortic calcification as a predictor of cardiovascular mortality.
Lancet 2: 1120-1122.
Witten ML; Wong SS; Sun NN; Keith I; Kweon C-B; Foster DE; Schauer JJ; Sherrill DL. (2005). Neurogenic responses in rat
lungs after nose-only exposure to diesel exhaust. Research report (Health Effects Institute): 1-47.
Wong CM; Atkinson RW; Anderson HR; Hedley AJ; Ma S; Chau PY; Lam TH. (2002a). A tale of two cities: effects of air
pollution on hospital admissions in Hong Kong and London compared. Environ Health Perspect 110: 67-77.
Wong SS; Sun NN; Keith I; Kweon C-B; Foster DE; Schauer James J; Witten ML. (2003). Tachykinin substance P signaling
involved in diesel exhaust-induced bronchopulmonary neurogenic inflammation in rats. Archives of toxicology 77: 638-
650.
Wong TW; Lau TS; Yu TS; Neller A; Wong SL; Tam W; Pang SW. (1999). Air pollution and hospital admissions for respiratory
and cardiovascular diseases in Hong Kong. Occupational and environmental medicine 56: 679-683.
Wong TW; Tam WS; Yu TS; Wong AH. (2002b). Associations between daily mortalities from respiratory and cardiovascular
diseases and air pollution in Hong Kong, China. Occupational and environmental medicine 59: 30-35.
Wongphatarakul V; Friedlander SK; Pinto JP. (1998). A comparative study of PM2.5 ambient aerosol chemical databases.
Environmental Science and Technology 32: 3926-3934.
Woodruff TJ; DarrowLA; Parker JD. (2008). Air pollution and postneonatal infant mortality in the United States, 1999-2002.
Environ Health Perspect 116: 110-115.
Woodruff TJ; Grillo J; Schoendorf KC. (1997). The relationship between selected causes of postneonatal infant mortality and
particulate air pollution in the United States. Environ Health Perspect 105.
Woodruff TJ; Parker JD; Schoendorf KC. (2006). Fine particulate matter (PM2 5) air pollution and selected causes of postneonatal
infant mortality in California. Environ Health Perspect 114: 786-790.
Wordley J; Walters S; Ayres JG. (1997). Short term variations in hospital admissions and mortality and particulate air pollution.
Occupational and environmental medicine 54: 108-116.
Wu CF; Delflno RJ; Floro JN; Samimi BS; Quintana PJ; Kleinman MT; Liu LJ. (2005). Evaluation and quality control of
personal nephelometers in indoor, outdoor and personal environments. J ExpoAnalEnvironEpidemiol 15: 99-110.
Wu J; M Winer A; J Delfino R. (2006). Exposure assessment of particulate matter air pollution before, during, and after the 2003
Southern California wildfires. Atmospheric Environment 40: 3333-3348.
Wu Y; Hao J; Fu L; Wang Z; Tang U. (2002). Vertical and horizontal profiles of airborne particulate matter near major roads in
Macao, China. Atmospheric Environment 36: 4907-4918.
Xia T; Korge P; Weiss JN; Li N; Venkatesen MI; Sioutas C; Nel A. (2004). Quinones and aromatic chemical compounds in
particulate matter induce mitochondrial dysfunction: implications for ultraflne particle toxicity. Environmental health
perspectives 112: 1347-1358.
Xiao GG; Nel AE; Loo JA. (2005). Nitrotyrosine-modified proteins and oxidative stress induced by diesel exhaust particles.
Electrophoresis 26: 280-292.
Xiaomin X; Zhen H; Jiasong W. (2006). The impact of urban street layout on local atmospheric environment. Building and
Environment 41: 1352-1363.
Xu GB; Yu CP. (1986). Effects of Age on Deposition of Inhaled Aerosols in the Human Lung. Aerosol Science and Technology
5: 349-357.
Xu J; DuBois D; Pitchford M; Green M; Etyemezian V. (2006). Attribution of sulfate aerosols in Federal Class I areas of the
western United States based on trajectory regression analysis. Atmospheric Environment 40: 3433-3447.
Xu J; Lee ET; Devereux RB; Umans JG; Bella M; Shara NM; Yeh J; Fabsitz RR; Howard BV. (2008). A longitudinal study of
risk factors for incident albuminuria in diabetic American Indians: the Strong Heart Study. Am J Kidney Dis 51: 415-424.
Yacobi NR; Phuleria HC; Demaio L; Liang CH; Peng CA; Sioutas C; Borok Z; Kim KJ; Crandall ED. (2007). Nanoparticle
effects on rat alveolar epithelial cell monolayer barrier properties. Toxicology in Vitro 21: 1373-1381.
Yang CY. (2008). Air pollution and hospital admissions for congestive heart failure in a subtropical city: Taipei, Taiwan. J
Toxicol Environ Health A 71: 1085-1090.
Yang CY; Chen CC; Chen CY; Kuo HW. (2007). Air pollution and hospital admissions for asthma in a subtropical city: Taipei,
Taiwan. J Toxicol Environ Health A 70: 111-117.
Yang CY; Chen YS; Yang CH; Ho SC. (2004a). Relationship between ambient air pollution and hospital admissions for
cardiovascular diseases in kaohsiung, taiwan. J Toxicol Environ Health A 67: 483-493.
Yang CY; Hsieh HJ; Tsai SS; Wu TN; Chiu HF. (2006). Correlation between air pollution and postneonatal mortality in a
subtropical city: Taipei, Taiwan. J Toxicol Environ Health A 69: 2033-2040.
Yang CY; Tseng YT; Chang CC. (2003a). Effects of air pollution on birth weight among children born between 1995 and 1997
in Kaohsiung, Taiwan. J Toxicol Environ Health A 66: 807-816.
December 2008
R-77
DRAFT-DO NOT QUOTE OR CITE
-------
Yang H; Li Q; Yu JZ. (2003b). Comparison of two methods for the determination of water-soluble organic carbon in atmospheric
particles. Atmospheric Environment 37: 865-870.
Yang Q; Chen Y; Krewski D; Shi Y; Burnett RT; McGrail KM. (2004b). Association between particulate air pollution and first
hospital admission for childhood respiratory illness in Vancouver, Canada. Archives of environmental health 59: 14-21.
Yang Z; Zhu L. (2007). Performance of the partition-limited model on predicting ryegrass uptake of polycyclic aromatic
hydrocarbons. Chemosphere 67: 402-409.
Yauk C; Polyzos A; Rowan-Carroll A; Somers CM; Godschalk RW; Van Schooten FJ; Berndt ML; Pogribny IP; Koturbash I;
Williams A; Douglas GR; Kovalchuk O. (2008). Germ-line mutations, DNA damage, and global hypermethylation in
mice exposed to particulate air pollution in an urban/industrial location. Proceedings of the National Academy of
Sciences of the United States of America 105: 605-610.
Yauk CL; Quinn JS. (1996). Multilocus DNA fingerprinting reveals high rate of heritable genetic mutation in herring gulls
nesting in an industrialized urban site. Proc Natl Acad Sci US A 93: 12137-12141.
Yeates DB; Gerrity TR; Garrard CS. (1981). Particle deposition and clearance in the bronchial tree. Annals of Biomedical
Engineering 9: 577-592.
Yeatts K; Svendson E; Creason J; Alexis N; Herbst M; Scott J; Kupper L; Williams R; Neas L; Cascio W. (2007). Coarse
particulate matter (PM10-2.5) affects heart rate variability, blood lipids, and circulating eosinophils in adults with asthma.
EHP 115: 709-714.
Yi SM; Ambs JL; Patashnick H; Rupprecht G; Hopke PK. (2004). Particle Collection Characteristics of a Prototype Electrostatic
Precipitator (ESP) for a Differential TEOM System. Aerosol Science and Technology 38: 46-51.
Yin XJ; Dong CC; Ma JYC; Antonini JM; Roberts JR; Stanley CF; Schafer R; Ma JKH. (2004a). Suppression of cell-mediated
immune responses to listeria infection by repeated exposure to diesel exhaust particles in brown Norway rats.
Toxicological sciences 77: 263-271.
Yin XJ; Ma JYC; Antonini JM; Castranova V; Ma JKH. (2004b). Roles of reactive oxygen species and heme oxygenase-1 in
modulation of alveolar macrophage-mediated pulmonary immune responses to Listeria monocytogenes by diesel exhaust
particles. Toxicological sciences 82: 143-153.
Ying Q; Kleeman MJ. (2006). Source contributions to the regional distribution of secondary particulate matter in California.
Atmospheric Environment 40: 736-752.
Yoshida S; Ono N; Tsukue N; Oshio S; Umeda T; Takano H; Takeda K. (2006a). In utero exposure to diesel exhaust increased
accessory reproductive gland weight and serum testosterone concentration in male mice. Environmental sciences 13: 139-
147.
Yoshida S; Takedab K. (2004). The effects of diesel exhaust on murine male reproductive function. Journal of health science 50:
210-214.
Yoshida S; Yoshida M; Sugawara I; Takeda K. (2006b). Mice strain differences in effects of fetal exposure to diesel exhaust gas
on male gonadal differentiation. Environmental sciences 13: 117-123.
Yu CP. (1985). Theories of electrostatic lung deposition of inhaled aerosols. Ann Occup Hyg 29: 219-227.
Yu H, et al. (2006). A review of measurement-based assessments of the aerosol direct radiative effect and forcing. Atmospheric
Chemistry and Physics 6: 613-666.
Yu H; Remer LA; Chin M; Bian H; Kleidman RG; Diehl T. (2008). A satellite-based assessment of transpacific transport of
pollution aerosol. J Geophys Res 113.
Yu IJ; Park JD; Park ES; Song KS; Han KT; Han JH; Chung YH; Choi BS; Chung KH; Cho MH. (2003). Manganese
Distribution in Brains of Sprague-Dawley Rats After 60 Days of Stainless Steel Welding-Fume Exposure.
Neurotoxicology 24: 777-785.
Yu JZ; Yang H; Zhang H; Lau AKH. (2004). Size distributions of water-soluble organic carbon in ambient aerosols and its size-
resolved thermal characteristics. Atmospheric Environment 38: 1061-1071.
Yu O; Sheppard L; Lumley T; Koenig JQ; Shapiro GG. (2000). Effects of ambient air pollution on symptoms of asthma in
Seattle-area children enrolled in the CAMP study. Environ Health Perspect 108: 1209-1214.
Yuangen Y; Campbell CD; Clark L; Camerson CM; Paterson E. (2006). Microbial indicators of heavy metal contamination in
urban and rural soils. Chemosphere 63: 1942-1952.
Yue W; Schneider A; Stolzel M; Ruckerl R; Cyrys J; Pan X; Zareba W; Koenig W; Wichmann HE; Peters A. (2007). Ambient
source-specific particles are associated with prolonged repolarization and increased levels of inflammation in male
coronary artery disease patients. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis 621: 50-60.
Yunker MB; Snowdon LR; MacDonald RW; Smith JN; Fowler MG; Skibo DN. (1996). Polycyclic aromatic hydrocarbon
composition and potential sources for sediment samples from the Beufort and Barents Seas. Environmental Science and
Technology 30: 1310-1320.
Zabel M; Klingenheben T; Franz MR; Hohnloser SH. (1998). Assessment of QT Dispersion for Prediction of Mortality or
Arrhythmic Events After Myocardial Infarction : Results of a Prospective, Long-term Follow-up Study. Circulation 97:
2543-2550.
Zakey AS; Solmon F; Giorgi F. (2006a). Development and testing of a desert dust module in a regional climate model. Atmos
Chem Phys Discuss 6: 1749-1792.
Zakey AS; Solmon F; Giorgi F. (2006b). Implementation and testing of a desert dust module in a regional climate model. Atmos
Chem Phys 6: 4687-4704.
December 2008
R-78
DRAFT-DO NOT QUOTE OR CITE
-------
Zanchi AC; Venturini CD; Saiki M; Nascimento Saldiva PH; Tannhauser Barros HM; Rhoden CR. (2008). Chronic nasal
instillation of residual-oil fly ash (ROFA) induces brain lipid peroxidation and behavioral changes in rats. Inhal Toxicol
20: 795-800.
Zanobetti A; Bind MAC; Schwartz J. (2008). Particulate air pollution and survival in a COPD cohort. Environmental Health 7:
48.
Zanobetti A; Canner MJ; Stone PH; Schwartz J; Sher D; Eagan-Bengston E; Gates KA; Hartley LH; Suh H; Gold DR. (2004).
Ambient Pollution and Blood Pressure in Cardiac Rehabilitation Patients. Circulation 110: 2184-2189.
Zanobetti A; Schwartz J. (2002). Cardiovascular damage by airborne particles: are diabetics more susceptible? Epidemiology 13:
588-592.
Zanobetti A; Schwartz J. (2003). Airborne particles and hospital admissions for heart and lung disease. In: Revised analyses of
time-series studies of air pollution and health. Special Report. Boston, MA: Health Effects Institute.
Zanobetti A; Schwartz J. (2005). The effect of particulate air pollution on emergency admissions for myocardial infarction: a
multicity case-crossover analysis. Environ Health Perspect 113: 978-982.
Zanobetti A; Schwartz J. (2006). Air pollution and emergency admissions in Boston, MA. Journal of epidemiology and
community health 60: 890-895.
Zanobetti A; Schwartz J. (2007). Particulate air pollution, progression, and survival after myocardial infarction. Environ Health
Perspect 115: 769-775.
Zanobetti A; Schwartz J; Samoli E; Gryparis A; Touloumi G; Peacock J; Anderson RH; Le Tertre A; Bobros J; Celko M; Goren
A; Forsberg B; Michelozzi P; Rabczenko D; Hoyos SP; Wichmann HE; Katsouyanni K. (2003). The temporal pattern of
respiratory and heart disease mortality in response to air pollution. Environ Health Perspect 111: 1188-1193.
Zappia G; Sabbioni C; Riontino C; Gobbi G; Favoni O. (1998). Exposure tests of building materials in urban atmosphere. The
Science of the total environment 224: 235-244.
Zechmeister HG. (1998). Annual growth of four pleurocarpous moss species and their applicability for biomonitoring heavy
metals. Environmental Monitoring and Assessment 52: 441-451.
Zechmeister HG; Hohenwallner D; Riss A; Hanus-Illnar A. (2003). Variation in heavy metal concentrations in the moss species
Abietinella abietina (Hedw.) Fleisch. according to sampling time, within site variability and increase in biomass. The
Science of the total environment 301: 55-65.
Zeger S; McDermott A; Dominici F; Samet J. (2007). Mortality in the medicare population and chronic exposure to fine
particulate air pollution. Johns Hopkins University, Dept of Biostatistics Working Papers 113.
Zeger SL; Thomas D; Dominici F; Samet JM; Schwartz J; Dockery D; Cohen A. (2000). Exposure Measurement Error in Time-
Series Studies of Air Pollution: Concepts and Consequences. Environmental health perspectives 108: 419-426.
Zeka A; Sullivan JR; Vokonas PS; Sparrow D; Schwartz J. (2006b). Inflammatory markers and particulate air pollution:
characterizing the pathway to disease. Int J Epidemiol 35: 1347-1354.
Zeka A; Zanobetti A; Schwartz J. (2005). Short term effects of particulate matter on cause specific mortality: effects of lags and
modification by city characteristics. Occupational and environmental medicine 62: 718-725.
Zeka A; Zanobetti A; Schwartz J. (2006a). Individual-level modifiers of the effects of particulate matter on daily mortality. Am J
Epidemiol 163: 849-859.
Zelikoff JT; Chen LC; Cohen MD; Fang K; Gordon T; Li Y; Nadziejko C; Schlesinger RB. (2003). Effects of inhaled ambient
particulate matter on pulmonary antimicrobial immune defense. Inhalation toxicology 15: 131-150.
Zelikoff JT; Schermerhorn KR; Fang K; Cohen MD; Schlesinger RB. (2002). A role for associated transition metals in the
immunotoxicity of inhaled ambient particulate matter. Environ Health Perspect 110 Suppl 5: 871-875.
Zeman KL; Bennett WD. (2006). Growth of the small airways and alveoli from childhood to the adult lung measured by aerosol-
derived airway morphometry. J Appl Physiol 100: 965-971.
Zender CS; Miller RL; Tegen I. (2004). Quantifying mineral dust mass budgets: Terminology, constraints, and current estimates.
Eos Trans Am Geophys Union 85: 509-512.
Zeng Y. (2006). A comprehensive particulate matter monitoring system and dosimetry-based ambient particulate matter
standards. J Air Waste Manag Assoc 56: 518-529.
Zhang J; Ghio AJ; Chang W; Kamdar O; Rosen GD; Upadhyay D. (2007a). Bim mediates mitochondria-regulated particulate
matter-induced apoptosis in alveolar epithelial cells. FEBS letters 581: 4148-4152.
Zhang KM; Wexler AS; Niemeier DA; Zhu YF; Hinds WC; Sioutas C. (2005a). Evolution of particle number distribution near
roadways. Part III: Traffic analysis and on-road size resolved particulate emission factors. Atmospheric Environment 39:
4155-4166.
Zhang Q; Jimenez JL; Canagaratna MR; Jayne JT; Worsnop DR. (2005b). Time- and size-resolved chemical composition of
submicron particles in Pittsburgh: Implications for aerosol sources and processes. Journal of Geophysical Research D:
Atmospheres 110: 1-19.
Zhang SH; Reddick RL; Piedrahita JA; Maeda N. (1992). Spontaneous hypercholesterolemia and arterial lesions in mice lacking
apolipoprotein E. Science 258: 468-471.
Zhang X; Kondragunta S; Schmidt C; Kogan F. (2008). Near real time monitoring of biomass burning particulate emissions
(PM2.5) across contiguous United States using multiple satellite instruments. Atmospheric Environment 42: 6959-6972.
Zhang Z; Che W; Liang Y; Wu M; Li N; Shu Y; Liu F; Wu D. (2007b). Comparison of cytotoxicity and genotoxicity induced by
the extracts of methanol and gasoline engine exhausts. Toxicol In Vitro 21: 1058-1065.
December 2008
R-79
DRAFT-DO NOT QUOTE OR CITE
-------
Zhao H; Barger MW; Ma JKH; Castranova V; Ma JYC. (2006a). Cooperation of the inducible nitric oxide synthase and
cytochrome P450 1A1 in mediating lung inflammation and mutagenicity induced by diesel exhaust particles.
Environmental health perspectives 114: 1253-1258.
Zhao W; Hopke PK; Norris G; Williams R; Paatero P. (2006b). Source apportionment and analysis on ambient and personal
exposure samples with a combined receptor model and an adaptive blank estimation strategy. Atmospheric Environment.
Zhao W; Rabinovitch N; Hopke PK; Gelfand EW. (2007). Use of an expanded receptor model for personal exposure analysis in
schoolchildren with asthma. Atmospheric Environment 41: 4084-4096.
Zhao X; Ju Y; Liu C; Li J; Huang M; Sun J; Wang T. (2008). Bronchial anatomy of left lung: a study of multi-detector row CT.
Surg Radiol Anat.
Zheng M; Cass GR; Schauer JJ; Edgerton ES. (2002). Source apportionment of PM2.5 in the Southeastern United States using
solvent-extractable organic compounds as tracers. Environ Sci Technol 36: 2361-2371.
Zheng M; Ke L; Edgerton ES; Schauer JJ; Dong M; Russell AG. (2006). Spatial distribution of carbonaceous aerosol in the
southeastern United States using molecular markers and carbon isotope data. J Geophys Res 111: D10S06.
Zhou M, et al. (2005). A normalized description of the direct effect of key aerosol types on solar radiation as estimated from
aerosol robotic network aerosols and moderate resolution imaging spectroradiometer albedos. Journal of Geophysical
Research 110, D19202, doi: 10.1029/2005JD005909.
Zhou Y; Levy JI. (2007). Factors influencing the spatial extent of mobile source air pollution impacts: a meta-analysis.
BMCPublic Health 7: 89.
Zhu K; Zhang J; Lioy PJ. (2007). Evaluation and comparison of continuous fine particulate matter monitors for measurement of
ambient aerosols. Journal of the Air & Waste Management Association 57: 1499-1506.
Zhu Y; Hinds WC; Kim S; Shen S; Sioutas C. (2002b). Study of ultrafine particles near a major highway with heavy-duty diesel
traffic. Atmospheric Environment 36: 4323-4335.
Zhu Y; Hinds WC; Kim S; Sioutas C. (2002a). Concentration and Size Distribution of Ultrafine Particles Near a Major Highway.
Journal of the Air & Waste Management Association 52: 1032-1042.
Zhu Y; Hinds WC; Shen S; Sioutas C. (2004). Seasonal trends of concentration and size distribution of ultrafine particles near
major highways in Los Angeles. Aerosol Science and Technology 38: 5-13.
Zhu Y; Kuhn T; Mayo P; Hinds WC. (2005). Comparison of Daytime and Nighttime Concentration Profiles and Size
Distributions of Ultrafine Particles near a Major Highway. Environmental science & technology 39: 2531-2536.
December 2008
R-80
DRAFT-DO NOT QUOTE OR CITE
------- |